Note from Andrey: this ep came out a week ago on RSS, but I was delayed posting it to youtube and therefore also Substack. My bad!

Our 238th episode with a summary and discussion of last week’s big AI news!

Recorded on 03/18/2026

Hosted by Andrey Kurenkov and Jeremie Harris

Feel free to email us your questions and feedback at andreyvkurenkov@gmail.com and/or hello@gladstone.ai

In this episode:

* OpenAI released GPT-5.4 mini and nano with 400k-token context windows, higher per-token prices but claimed token-efficiency gains in Codex; nano is API-only and pitched for high-volume classification/data extraction despite a major price increase.

* Mistral open-sourced the Small 4 model family (MoE, 119B total/6B active) combining reasoning, multimodal, and coding-agent capabilities, and announced Forge to help businesses train or post-train custom models.

* Agent “operating system” competition intensified with Meta’s acquired Manus launching a local Mac agent, Nvidia announcing NeMo/“Open Shell” sandboxed agent runtime, and Nvidia also unveiling DLSS 5 plus major hardware forecasts including Groq LPU integration.

* Business and safety updates included OpenAI shifting focus toward productivity/enterprise amid competition, Microsoft reorganizing Copilot and frontier-model efforts, Meta delaying its next model, China-linked ByteDance deploying large Nvidia clusters abroad, and new safety work on steganography, chain-of-thought faithfulness, fine-tuning defenses, cyber-attack evals, and constitution/spec compliance.

A thank you to our current sponsors:

• Box - visit Box.com/AI to learn more

• ODSC AI - go to odsc.ai/east and use promo code LWAI for an additional 15% off your pass to ODSC AI East 2026.

• Factor - head to factormeals.com/lwai50off and use code lwai50off to get 50 percent off and free breakfast for a year

Timestamps:

• (00:00:10) Intro / Banter

• (00:01:56) News Preview

• Tools & Apps

• (00:02:39) OpenAI ships GPT-5.4 mini and nano, faster and more capable but up to 4x pricier

• (00:08:04) Mistral’s new Small 4 model punches above its weight with 128 expert modules

• (00:14:03) Meta’s Manus launches ‘My Computer’ to turn your Mac into an AI agent - 9to5Mac

• (00:17:57) NVIDIA Announces NemoClaw for the OpenClaw Community | NVIDIA Newsroom + Nvidia boosts knowledge work with Open Agent Development Platform

• (00:24:09) DLSS 5 looks like a real-time generative AI filter for video games | The Verge

• (00:26:36) OpenAI to Launch ChatGPT ‘Adult Mode’ Despite Warnings From Its Own Advisers - CNET

• Applications & Business

• (00:33:46) OpenAI Reportedly Pivoting to a Focus on Business and Productivity Only

• (00:41:25) Nvidia GTC 2026: CEO Jensen Huang sees $1 trillion in orders for Blackwell and Vera Rubin through ’27

• (00:45:44) Mistral launches Forge to help enterprises build their own AI models

• (00:54:17) China’s ByteDance gets access to top Nvidia AI chips, WSJ reports

• (00:57:57) Meta Delays Rollout of New A.I. Model After Performance Concerns

• (01:02:50) Microsoft Shakes Up AI Division As Copilot Falls Behind Google and OpenAI

• Policy & Safety

• (01:07:26) A Decision-Theoretic Formalisation of Steganography With Applications to LLM Monitoring

• (01:13:09) Reasoning Theater: Disentangling Model Beliefs from Chain-of-Thought

• (01:18:29) In-Training Defenses against Emergent Misalignment in Language Models

• (01:23:07) How do frontier AI agents perform in multi-step cyber-attack scenarios?

• (01:25:20) Eval awareness in Claude Opus 4.6’s BrowseComp performance

• (01:29:49) Introducing Bloom: an open source tool for automated behavioral evaluations

• (01:32:26) How well do models follow their constitutions?

• (01:37:11) Nvidia’s H200 License Stirs Security Concern Among Top Democrats

• Research & Advancements

• (01:40:050) [2603.15031] Attention Residuals

• (01:47:11) Mamba-3: Improved Sequence Modeling using State Space Principles

Related:

• Nvidia’s DLSS 5 uses generative AI to boost photorealism in video games, with ambitions beyond gaming

Summary: Nvidia unveiled DLSS 5, calling it a “GPT moment for graphics” that blends traditional 3D rendering with generative AI to boost photorealism in real time up to 4K. Unlike prior DLSS upscalers, DLSS 5’s end-to-end AI model analyzes a single frame’s scene semantics—characters, hair, fabric, translucent skin—and lighting conditions (front-lit, back-lit, overcast) to generate new detail. Early examples from Resident Evil Requiem, Starfield, Hogwarts Legacy, and EA Sports FC show sharper lighting and shadows but also noticeable alterations to character materials and faces—e.g., Requiem’s Grace Ashcroft appearing with fuller lips and heavy eyeshadow, and Starfield models looking stage-lit and hyper-sharpened.

Nvidia says artistic intent is preserved by anchoring outputs with per-frame color and motion vectors and giving developers granular controls over intensity, color grading, blending, contrast, saturation, gamma, per-object masking, and exclusion zones; still, some developers like Mike Bithell criticized the look as removing art direction. On the technical side, DLSS 5 fuses structured 3D data (the “ground truth” of the virtual scene) with generative, probabilistic models that can predict and fill in image elements instead of rendering everything from scratch, aiming to deliver “beautiful, amazing, as well as controllable” results with less compute.

DLSS 5 is slated for release this fall with initial support confirmed for titles such as Starfield, Resident Evil Requiem, Hogwarts Legacy, EA Sports FC, The Elder Scrolls VI: Oblivion remake, and Assassin’s Creed Shadows.

Editor’s Take: The initial reaction to this seemed heavily negative to me, which is a shame considering this must be a massive technical achievement. Nvidia messed up by not making artist control front and center, leaving the impression that this is just an ‘AI filter’ that will just make games look worse. Still, personally i’m excited to see how this looks in practice!

OpenAI Reportedly Pivoting to a Focus on Business and Productivity Only

Related:

• OpenAI is planning a desktop ‘superapp’

• Panicked OpenAI Execs Cutting Projects as Walls Close In

• OpenAI shifts to coding and enterprise as Anthropic pulls ahead

Summary: OpenAI is pivoting hard toward business and productivity, with Chief of Applications Fidji Simo announcing plans to merge ChatGPT, the Codex coding platform, and the Atlas browser into a single desktop “superapp.” The plan reverses a product strategy from last year that left the company scattered, with multiple individual apps that drew an uneven response from users and pulled internal attention in different directions.

In an internal memo, Simo wrote: “We realized we were spreading our efforts across too many apps and stacks, and that we need to simplify our efforts. That fragmentation has been slowing us down and making it harder to hit the quality bar we want.”

The centerpiece of the combined app will be “agentic” AI — tools designed to run independently on a computer and handle tasks ranging from coding to data analysis. In the near term, Codex will be expanded to handle productivity work beyond coding, with ChatGPT and Atlas brought into the unified app in later phases. The mobile ChatGPT app will remain unchanged.

The urgency is hard to miss. Anthropic’s portion of enterprise AI spending has climbed to 40% while OpenAI’s share of the same market fell from roughly half to about 27%. At an all-hands meeting, Simo reportedly told employees they couldn’t afford to be distracted by “side quests” given Anthropic’s rapid success winning over enterprise and coding customers.

Editor’s Take: I’d say it’s fair to call this unsurprising —* OpenAI’s various bets (Sora, Atlas, Prism) seemed not to have paid off much, and their lack of focus on Codex no doubt hurt their ability to compete with Anthropic’s Claude Code / Cowork. As a big fan of Claude Code, i’m happy there is healthy competition in the space to make sure both Codex and Claude improve. Aside from that, i’m curious to see what this ‘Superapp’ might look like, assuming it ever actually does release.

*For the record: I added that em-dash, not AI!

Meta’s Manus Launches Desktop App With AI Agent for Tasks Across Files, Apps

Related:

• Meta’s Manus launches ‘My Computer’ to turn your Mac into an AI agent

Summary: Meta’s newly acquired startup Manus released a desktop app for Mac (Apple Silicon) and Windows that brings its agentic system “My Computer” onto local machines. The app presents a chatbot-style interface with a central prompt and options to attach files or folders, then executes command line instructions (CLI) in the system terminal to carry out tasks. Capabilities include reading, analyzing, and editing local files; launching and controlling local applications; and performing bulk operations like sorting thousands of photos into categorized subfolders or renaming large batches of invoices. It can convert file formats, build simple apps, and even use a local GPU to train a machine learning model or run a large language model for inference.

The tool also supports remote actions and Google app integrations so users can, for example, fetch a desktop file and have the agent email it to a client while away. Each folder added for automation triggers a permission prompt with Allow, Always Allow, or Cancel, and the app is available with a limited free plan and paid tiers starting at $20/month ($17 billed annually). The release follows attention on similar AI agents like OpenClaw and Perplexity’s Personal Computer, with experts warning about privacy and security risks from agents that execute system-level commands. Manus began in China and moved HQ to Singapore; Chinese authorities are reportedly reviewing the legality of its acquisition by Meta, which previously offered only cloud-based services before this on-device expansion.

Editor’s Take: Spearking of Codex and Cowork, I guess Meta is jumping in on that bandwagon as well? A bit of a weird move, even if it makes sense for Manus. As with LLM chatbots back in 2023, deep research-type agents in 2024, and reasoning models in 2025, it seems like all the big players in AI are investing in the new hot (deservingly so) trend and generally converging to very similar offerings.

MiniMax M2.7 Testing Shows Benchmark Wins & Major Cost Savings

Related:

• MiniMax launches M2.7 model on MiniMax Agent and APIs

• Cursor’s Composer 2 beats Opus 4.6 on coding benchmarks at a fraction of the price

Summary: MiniMax’s new M2.7 model posts strong agentic performance at unusually low cost, with benchmark scores of 56.22% on Swaybench/SWE-Pro, 55.6% on VIBE-Pro, and 57% on Terminal-Bench 2. It emphasizes autonomous self-improvement, running 100+ self-training cycles via agent harnesses and reinforcement learning that the company says yield a 30% capability lift, plus 97% skill adherence across 40+ complex skills and a 24K context window. The release is live on MiniMax Agent and APIs, supports multi-agent collaboration, autonomous debugging, research agent harnesses, and more.

Pricing is a standout: as low as $0.30 per million input tokens and $120 per million output tokens, with an optional fast mode at 2x price, positioning M2.7 as up to 50x cheaper than Opus 4.6 while claiming wins over Gemini 3.1 Pro and competitive Terminal-Bench 2 performance for enterprise workflows in finance, ML pipelines, game dev, and dynamic web UIs.

Relatedly, Cursor’s Composer 2 arrives as a parallel push on affordable agentic coding, beating Opus 4.6 on Terminal-Bench 2.0 with 61.7% (vs. 58.0%) at $0.5/$2.5 per million input/output tokens (fast mode $1.5/$7.5), though still below GPT-5.4’s 75.1%. A key technical novelty is “self-summarization,” a compaction-in-the-loop RL method that trains the model to pause on token-length triggers and compress its own action history to ~1,000 tokens from 5,000+, with rewards spanning the entire trajectory; Cursor reports 50% fewer compaction errors and stronger long-horizon task handling.

Editor’s Take: MiniMax and Chinese labs in general continue to impress with their even improving models, which at this point are more than capable of taking care of a lot of stuff that only Western closed source models used to be capable of. Cursor got some flak for training Composer 2 on top of Moonshot AI’s Kimi, which is quite silly - starting with already strong open source models and training them further should by now be the no-brainer move for any AI company whose primary business isn’t already frontier model development.

Other News

Tools

OpenAI ships GPT-5.4 mini and nano, faster and more capable but up to 4x pricier. These smaller models nearly match full GPT-5.4 on coding, reasoning, and multimodal benchmarks while running faster and offering a 400k-token context, but they come with input/output pricing up to four times higher than the previous mini and nano models.

Mistral bets on ‘build-your-own AI’ as it takes on OpenAI, Anthropic in the enterprise. The new Forge platform lets companies train and deploy custom models from scratch on their own data (with Mistral guidance and embedded engineers), targeting enterprise needs like language, compliance, and domain-specific performance.

Mistral’s new Small 4 model punches above its weight with 128 expert modules. It routes queries through 128 expert modules but activates just four per request to keep responses fast and efficient, lets users trade off speed versus thoroughness, and is available under Apache 2.0 on Hugging Face, Mistral’s API, and Nvidia platforms.

Nvidia Debuts Platform for Enterprise AI Agents. The offering provides security, privacy controls, and policy enforcement so companies can deploy OpenClaw-style autonomous AI assistants while limiting data access, controlling actions, and enabling audits.

NVIDIA Announces NemoClaw for the OpenClaw Community. NemoClaw installs OpenShell and Nemotron models with a single command to provide sandboxed, policy-driven privacy and security controls that let always-on OpenClaw agents run locally or leverage cloud models on NVIDIA RTX and DGX systems.

The Gemini-powered features in Google Workspace that are worth using. Google is rolling out practical tools across Docs, Gmail, Sheets, Slides, Drive, Meet, Calendar, Chat, Vids, and Forms—like summarization, draft generation, data extraction, automatic meeting notes, scheduling help, and content formatting—that speed up everyday workflows and information management.

Microsoft launched a second-generation version of its AI image model.. The update improves image quality and consistency and is being rolled out to Microsoft’s image-generation features and developer APIs.

Adobe’s AI image generator can now be trained on your own art. Users can now train private Firefly Custom Models on their own assets to produce consistent character designs, illustrations, and photos at scale while preventing opted-out content from being used.

Google tests voice cloning on AI Studio powered by Gemini. A hidden “Create Your Voice” option and related UI hints suggest Google is building native voice-cloning into AI Studio (tied to Gemini 2.5 Flash now) that would let developers generate synthetic voices from user-provided samples, alongside upcoming GitHub repo import and other developer-focused integrations.

Perplexity launches consumer-focused AI health tool. The new tool aims to combine EHR and wearable data to provide consumer health insights, entering a crowded space of AI-driven health assistants.

Business

Waymo hits 170 million miles while avoiding serious mayhem. Waymo reports its fleet has logged over 170 million miles with far fewer serious-injury crashes than human drivers but faces scrutiny from safety advocates over how it frames its data, incidents involving pedestrians and emergency vehicles, and the limited scale of its operations.

OpenAI expands government footprint with AWS deal, report says. The agreement lets AWS distribute OpenAI’s models through its GovCloud and Classified Regions for Secret and Top Secret workloads while OpenAI retains control over which models are offered and can impose deployment-specific safeguards.

Microsoft may take legal action over Amazon-OpenAI deal. Microsoft says it is reviewing whether Amazon Web Services hosting OpenAI’s new commercial product “Frontier” would violate an exclusivity clause that requires OpenAI’s models to run on Azure.

Microsoft Shakes Up AI Division As Copilot Falls Behind Google and OpenAI. The reorganization shifts Suleyman to focus solely on developing Microsoft’s own frontier language models while Jacob Andreou takes charge of unifying and growing the Copilot consumer and commercial products to reduce reliance on OpenAI and address weak user adoption.

Mistral AI makes enterprise push with two new launches. The new offerings include Mistral Small 4 — a 119B-parameter hybrid multimodal model claimed to improve reasoning, coding, and throughput versus its predecessor — and Mistral Forge, a platform that lets enterprises train custom models on proprietary data.

Meta says its AI moderation systems will replace contractors over the next few years.. The company plans to roll out an AI support assistant across Facebook and Instagram that it says will reduce reliance on third-party moderation contractors over the next few years.

OpenAI to acquire developer tooling startup Astral in boost for Codex team. The small team’s engineers will join OpenAI to work on its Codex coding assistant, bolstering the company’s developer tooling amid rapid user growth and ongoing acquisition activity.

Research

V-JEPA 2.1: Unlocking Dense Features in Video Self-Supervised Learning. The authors train a unified image‑and‑video JEPA with a dense predictive loss applied to all tokens (and deep hierarchical supervision), producing higher‑quality spatio‑temporal dense features that improve forecasting, segmentation, depth, and robot planning performance.

Attention Residuals. This work replaces fixed, equal-weight residual aggregation with a content-dependent softmax attention over previous layer outputs (plus a blockwise, memory-efficient variant) to prevent hidden-state dilution, improve depth‑wise signal/gradient balance, and boost downstream performance in large-scale LLMs.

SWE-Skills-Bench: Do Agent Skills Actually Help in Real-World Software Engineering?. The benchmark finds that injecting off-the-shelf SWE agent skills yields minimal average improvement (+1.2% pass-rate), with most skills producing no benefit, a few specialized skills offering up to +30% gains, and some causing negative interference when conventions mismatch project context.

GradMem: Learning to Write Context into Memory with Test-Time Gradient Descent. The method optimizes a small set of writable memory‑token embeddings at test time using a self‑supervised reconstruction loss—keeping model weights fixed—to compactly store context information with a few gradient steps, yielding higher memory capacity than forward‑only encoding and transferring to some natural‑language tasks.

Delightful Policy Gradient. The proposed “Delightful Policy Gradient” multiplies each sample’s policy‑gradient term by a sigmoid of (advantage × action surprisal) with a fixed temperature to reduce harmful updates from unlikely or already‑solved actions and shift gradients closer to supervised cross‑entropy.

Concerns

OpenAI’s own mental health experts unanimously opposed “naughty” ChatGPT launch. Council members warned that AI-generated erotica could foster unhealthy emotional dependence and enable minors to access sexual chats, raising concerns about risks like users being encouraged toward self-harm.

Encyclopedia Britannica and Merriam-Webster Sue OpenAI. Britannica and Merriam‑Webster allege OpenAI trained ChatGPT on their content without permission and that the chatbot reproduces or closely paraphrases that copyrighted material, reducing traffic to the original sites.

Our 237th episode with a summary and discussion of last week’s big AI news!

Recorded on 03/13/2026

Hosted by Andrey Kurenkov and Jeremie Harris

Feel free to email us your questions and feedback at andreyvkurenkov@gmail.com and/or hello@gladstone.ai

In this episode:

* Perplexity announced “Personal Computer,” a local Mac-based AI agent positioned as a safer alternative to OpenAI’s computer-use agents, while Anthropic added GitHub PR code review pricing reviews at $15–$25 and Cursor launched trigger-based “Automations” for always-on coding agents.

* ChatGPT introduced interactive math/science visuals and Anthropic added in-chat interactive charts/diagrams; Nvidia released open weights for its 120B-parameter Natron Free Super hybrid Transformer–Mamba latent-MoE model trained natively at 4-bit for Blackwell GPUs.

* Nvidia halted H200 production for China amid customs blocks and domestic chip pressure; xAI saw major co-founder departures; Anthropic previewed a Claude Marketplace for enterprise procurement; Yann LeCun’s aMI raised $1.3B; humanoid robot maker Sanctuary reached a $1.15B valuation.

* Anthropic sued the Pentagon over a “supply chain risk” designation as memos ordered removal within 180 days; research covered models resisting activation steering, limits of chain-of-thought control, inference-scaling boosting cyber-task success, low-probability risky actions, weaknesses in SWE-bench, multimodal pretraining, long-context RNN memory caching, context-parallel training efficiency, RL for CUDA kernel optimization, and latent introspection detecting concept injection.

A thank you to our current sponsors:

• Box - visit Box.com/AI to learn more

• ODSC AI - go to odsc.ai/east and use promo code LWAI for an additional 15% off your pass to ODSC AI East 2026.

• Factor - head to factormeals.com/lwai50off and use code lwai50off to get 50 percent off and free breakfast for a year

Timestamps:

• (00:00:10) Intro / Banter

• (00:01:23) Response to listener comments

• Tools & Apps

• (00:02:06) Perplexity’s Personal Computer turns your spare Mac into an AI agent | The Verge

• (00:04:22) Anthropic launches code review tool to check flood of AI-generated code | TechCrunch

• (00:08:08 ) Cursor is rolling out a new kind of agentic coding tool | TechCrunch

• (00:11:14) ChatGPT can now create interactive visuals to help you understand math and science concepts | TechCrunch

• (00:11:56) Anthropic’s Claude AI can respond with charts, diagrams, and other visuals now | The Verge
  

• Projects & Open Source

• (00:13:54) Introducing Nemotron 3 Super: An Open Hybrid Mamba-Transformer MoE for Agentic Reasoning | NVIDIA Technical Blog
  

• Applications & Business

• (00:21:22) Nvidia halts H200 production as China backs Huawei AI chips

• (00:28:33) Another XAI Cofounder Has Left, and Another Says He’s Leaving. - Business Insider

• (00:34:04) Anthropic’s Claude Marketplace allows customers to buy third-party cloud services | TechRadar

• (00:37:57) Yann LeCun’s AMI Labs raises $1.03 billion to build world models | TechCrunch

• (00:44:52) Humanoid robotics maker Sunday reaches $1.15B valuation to build household robots | TechCrunch
  

• Policy & Safety

• (00:46:09) Anthropic Sues Department of Defense Over ‘Supply Chain Risk’ Label - The New York Times + Google and OpenAI Just Filed a Legal Brief in Support of Anthropic

• (00:53:24) Internal Pentagon memo orders military commanders to remove Anthropic AI technology from key systems - CBS News

• (00:58:15) Endogenous Resistance to Activation Steering in Language Models

• (01:06:27) Reasoning Models Struggle to Control their Chains of Thought

• (01:09:52) ‘It means missile defence on datacentres’: drone strikes raise doubts over Gulf as AI superpower

• (01:14:57) Evidence for inference scaling in AI cyber tasks: Increased evaluation budgets reveal higher success rates

• (01:18:24) Frontier Models Can Take Actions at Low Probabilities
  

• Research & Advancements

• (01:24:20) Research note: Many SWE-bench-Passing PRs Would Not Be Merged into Main

• (01:28:26) [2603.03276] Beyond Language Modeling: An Exploration of Multimodal Pretraining

• (01:40:09) Memory Caching: RNNs with Growing Memory

• (01:48:47) Untied Ulysses: Memory-Efficient Context Parallelism via Headwise Chunking

• (01:58:41) CUDA Agent: Large-Scale Agentic RL for High-Performance CUDA Kernel Generation

• (02:08:57) Latent Introspection: Models Can Detect Prior Concept Injections

• (02:16:45) Physics of RL: Toy scaling laws for the emergence of reward-seeking

Related:

• OpenAI and Google Workers File Amicus Brief in Support of Anthropic Against the US Government

• Internal Pentagon memo orders military commanders to remove Anthropic AI technology from key systems

Summary: Anthropic filed two lawsuits—one in the Northern District of California and one in the D.C. Circuit—arguing the Pentagon’s new “supply‑chain risk to national security” designation and a White House-directed government-wide ban are unlawful retaliation after negotiations over usage limits for Claude collapsed. The company contends the unprecedented SCR label for a U.S. firm is already jeopardizing hundreds of millions of dollars and violates required procedures and presidential authority; it is seeking a temporary restraining order to continue work with military partners.

An internal DoD memo dated March 6 ordered all commanders to remove Anthropic AI from Defense systems and networks within 180 days, including nuclear, missile defense, cyber warfare systems, and any contractor work, with narrow exemptions requiring CIO approval and risk‑mitigation plans. Pentagon officials say they need Claude for “all lawful purposes,” while Anthropic’s red lines sought to bar mass domestic surveillance and fully autonomous lethal weapons; DoD has reportedly used Claude on classified networks for intelligence synthesis, targeting recommendations, and battle simulations in partnership with Palantir.

Support and backlash quickly mounted across the AI sector and government. More than 30 employees from OpenAI and Google, including Google DeepMind’s Jeff Dean and researchers from both labs, filed an amicus brief supporting Anthropic’s TRO, warning that blacklisting introduces unpredictability, chills debate on frontier AI risks, and could have broader competitiveness consequences; Sam Altman publicly called enforcing the SCR “very bad” even as OpenAI inked its own Pentagon deal.

The White House stated the administration will not let a “woke AI company’s terms of service” constrain the military, while DoD declined to comment on litigation.

Editor’s Take: Last week I wrote “Still, it appears likely that this story is not yet over”, and so it is. Now that things have gone to the court it may well be that things will drag on for months, and the situation will quiet down for a time. There’s a lot to be said about this whole affair, and you can hear more about it in our latest podcast episode.

‘Not built right the first time’ — Musk’s xAI is starting over again, again

Related:

• The XAI Exodus: Two More Cofounders Leave As Musk Says He’s Rebuilding

• Musk’s xAI wins permit for datacenter’s makeshift power plant despite backlash

• XAI’s Macrohard project stalls as Tesla ramps up a similar AI agent effort

Summary: Elon Musk says xAI is being “rebuilt from the foundations up,” as the startup undergoes a sweeping reorg and leadership exodus while lagging rivals in AI coding tools. Two more cofounders, Zihang Dai and Guodong Zhang (who led Grok Code and Grok Imagine), departed this week, leaving only Manuel Kroiss and Ross Nordeen from the original 11 founders; earlier exits included Toby Pohlen, Jimmy Ba, Tony Wu, and Greg Yang.

Musk acknowledged “Grok is currently behind in coding” versus Anthropic’s Claude Code and OpenAI’s Codex, held an all‑hands to course‑correct, and predicted catch‑up by mid‑year. xAI has shed dozens of employees since January, brought in SpaceX/Tesla execs to evaluate and cut staff, and is now combing through previously rejected applicants; it also hired Cursor’s Andrew Milich and Jason Ginsberg to bolster product engineering for coding assistants. Meanwhile, Grok Imagine (image/video generation) and key initiatives have faced cuts as the company prioritizes revenue‑driving coding tools and tries to show traction now that xAI sits under SpaceX ahead of a potential IPO.

Macrohard, xAI’s “AI white‑collar worker” agent meant to perform end‑to‑end computer tasks, has stalled after leadership churn and a paused data effort involving 600 contractors asked to screen‑record workflows; many engineers left or moved teams, and job listings for Macrohard have disappeared. Musk now frames Macrohard as a joint xAI–Tesla project alongside Tesla’s “Digital Optimus,” with Grok acting as the high‑level planner and the Tesla agent handling continuous, real‑time control—an approach modeled on Tesla Full Self‑Driving’s video‑based pipeline rather than screenshot‑by‑screenshot agents.

Editor’s Take: Elon Musk said “xAI was not built right first time around, so is being rebuilt from the foundations up”, and that ‘from the foundations up’ bit certainly appears to be true. Re-focusing away from Grok Imagine towards competing in the AI coder space makes sense strategically (OpenAI seems to have similarly prioritized Codex in recent months), but given xAI’s late start and current chaos it’s hard to imagine them catching up.

Cascade of A.I. Fakes About War With Iran Causes Chaos Online

Summary: Over 110 unique AI‑generated fakes about the new Iran war circulated across X, TikTok, Facebook, and private messaging apps in just two weeks, amassing millions of views, according to the New York Times. These items covered active combat, preparation, destruction, and propaganda: 37 pieces falsely depicting ongoing warfare, 5 on war preparations, 8 on destruction, 5 showing crying soldiers, 43 memes or overt AI content, and 13 other fabricated items.

To verify fakes, reporters combined visual tells (nonexistent buildings, garbled text, physics‑defying motion), invisible watermarks, multiple AI‑detector tools, and cross‑checks against reliable reporting. Experts say this wave outpaces previous conflicts due to more capable, low‑cost generative tools (including video models like Sora), multi‑front hostilities, and pro‑Iran narratives emphasizing military prowess and regional devastation.

Several clips went intensely viral, such as a balcony “Tel Aviv under missile barrage” video with an inserted Israeli flag—a common artifact when prompts mention Israel—and spectacular, Hollywood‑like scenes with mushroom clouds, hypersonic streaks, and sonic booms not seen in genuine battlefield footage. AI fabrications also fueled misinformation around the alleged attack on the U.S.S. Abraham Lincoln, with numerous bogus clips showing carriers ablaze despite U.S. statements that the ship was unharmed. Some content was openly propagandistic, including dramatized short films of the Shajarah Tayyebeh school strike and flattering or dehumanizing leader portrayals.

Platform responses remain thin: watermarking is easily stripped, few posts bore labels, and X’s new rule only demonetizes unlabeled “armed conflict” AI posts for 90 days.

Editor’s Take: Concern over ‘DeepFakes’ was all the rage in the late 2010s when generative AI for images was just starting to get good. Fears that DeepFakes’ harms would spread quickly proved unfounded (though ‘undressing’ apps did do real harm), but over the past several years we’ve seen a gradual growth in the impact of AI-generated images in all sorts of ways (scams, propaganda, brainrot content, false media of wars). The economic and psychological harm it brings to people is unqestionable and saddening. Are these false depicisions of war as harmful? I don’t know, but the fact they contribute to this being ‘the era of post-truth’ frustrates me immensely.

You can now ask Google Maps ‘complex, real-world questions’ — and Gemini will answer

Summary: Google is rolling out Ask Maps, an AI‑powered conversational search in Google Maps that uses Gemini to handle “complex, real‑world questions.” Users can describe plans in natural language (e.g., “find a vegetarian spot between Midtown East and my office with a cozy aesthetic and a table for four at 7pm”), and Maps will parse reviews, photos, and busyness data to surface tailored options—then book a table with a tap.

Personalization draws only from Google Maps and relevant prior searches tied to saved or favorited places; it does not use data from other Google apps like Gmail, according to Google. Paid placements currently do not influence Ask Maps recommendations. The feature launches this week in the US and India on Android and iOS, with desktop support coming soon.

A major navigation overhaul is arriving as well. Google is introducing Immersive Navigation, which it calls the biggest Maps upgrade in over a decade. The interface adds refreshed colors, detailed 3D buildings, elevated roadways, realistic terrain and greenery, plus dynamic camera zoom that shifts to highlight upcoming maneuvers. It explicitly calls out lanes, crosswalks, traffic lights, and stop signs when relevant, and explains route choices using live traffic plus user‑reported construction, crashes, and hazards; it can also provide parking info and walking directions after arrival.

Immersive Navigation begins rolling out in the US next week on iOS and Android, and will be available on Apple CarPlay, Android Auto, and vehicles with Google built in.

Editor’s Take: Finally, Google brings some Gemini greatness to its actual product line. Despite Gemini’s technical excellenge and Deepmind’s relentless pace over the past year, it’s felt like Google has kept up as far as actually improving their consumer offerings. Unlike many bolted-on barely useful AI features, this appears legimitately cool, and a step in the right direction.

Other News

Tools

Cursor is rolling out a new kind of agentic coding tool. Called Automations, the feature launches and manages coding agents automatically (triggered by code changes, Slack messages, timers, or incidents) so engineers only intervene at key review points.

Anthropic launches code review tool to check flood of AI-generated code. Integrated with GitHub, the tool automatically analyzes and comments on pull requests—focusing on logical errors, flagging severity levels, offering step‑by‑step explanations and suggested fixes, and allowing customization and light security checks for enterprise teams.

ChatGPT can now create interactive visuals to help you understand math and science concepts. Users get over 70 manipulable modules for math and science topics, where adjusting variables updates formulas, diagrams, and results in real time directly within ChatGPT.

Perplexity’s Personal Computer turns your spare Mac into an AI agent. The tool runs continuously on a dedicated Mac on your local network, giving the agent full access to your files and apps, remote control from any device, and safety features like an audit trail, action approvals, and a kill switch, with early access available via a waitlist.

Introducing Nemotron 3 Super: An Open Hybrid Mamba-Transformer MoE for Agentic Reasoning. NVIDIA is releasing an open 120B‑total (12B active) hybrid Mamba‑Transformer MoE trained natively in NVFP4 with latent MoE, multi‑token prediction, a 1M‑token context window, and RL‑tuned workflows to improve throughput, long‑context reasoning, and deployment efficiency for multi‑agent and agentic tasks.

You can (sort of) block Grok from editing your uploaded photos. A new toggle prevents Grok from being tagged to edit an uploaded image, but it doesn’t stop other workarounds or broader abuses of its image‑generation features.

Anthropic’s Claude AI can respond with charts, diagrams, and other visuals now. The update lets Claude automatically insert interactive charts, diagrams, and other visualizations directly into chats (or generate them on request); images are editable or clickable for more information and change as the conversation evolves.

Meta Is Developing 4 New Chips to Power Its AI and Recommendation Systems. Built with Broadcom on RISC‑V and TSMC fabrication, the MTIA 300 is already in production for ranking model training, with three upcoming inference‑focused chips (MTIA 400/450/500) planned to ship between early and late 2027 as part of Meta’s strategy to supplement, not replace, purchases from Nvidia, AMD, and others.

Gemini’s task automation is here and it’s wild. Gemini can now take actions inside apps—like ordering rides or food—by following prompts, asking clarifying questions, and performing steps in a virtual window while pausing for user confirmation before finalizing.

Business

Zoox starts mapping Dallas and Phoenix for its robotaxis. Mapping and initial SUV‑based testing will let Zoox collect driving data, open depots and a Scottsdale command center, and begin local trials before deploying its purpose‑built robotaxis pending federal and local approvals.

Anthropic’s Claude Marketplace allows customers to buy third-party cloud services. Enterprise customers can use existing Anthropic spending commitments to purchase third‑party cloud and AI services (starting with six partners) in one consolidated billing portal while Anthropic takes no cut.

Yann LeCun’s AMI Labs raises $1.03B to build world models. The funding will bankroll AMI Labs’ multi‑year effort to develop “world models” that learn from real‑world data (using approaches like JEPA) and to partner with companies such as Nabla for early health‑focused testing, while prioritizing open research over near‑term revenue.

Qualcomm, Wayve partner to accelerate AI-powered self-driving system rollout. The partnership combines Wayve’s data‑driven AI Driver software with Qualcomm’s Snapdragon Ride chips and safety stack to offer carmakers a scalable, standardized platform that shortens integration complexity and supports features from hands‑off to advanced “eyes‑off” driving as regulations permit.

Nissan, Uber reportedly finalizing deal for Wayve-powered robotaxi rollout. Demonstrated in an Ariya electric crossover, the system will arrive in Nissan’s 2027 fiscal year and use artificial intelligence to power its next‑generation ProPilot driving features.

Anthropic is launching a new think tank amid Pentagon blacklist fight. The company will combine three existing teams into a roughly 30‑person research unit led by Jack Clark to study large‑scale societal, economic, and safety implications of powerful AI, expanding into new projects and staffing despite Anthropic’s ongoing dispute with the Pentagon.

Humanoid robotics maker Sunday reaches $1.15B valuation to build household robots. The Series B will fund development of Sunday’s humanoid household robot Memo, which the startup says will assist with chores like laundry and clearing the table as it scales toward production and addresses longstanding challenges in robot manipulation.

Anthropic Pours $100 Million Into Claude Partner Network In Channel Push. New funding will support partner training, go‑to‑market efforts, direct partner funding for deployments and co‑marketing, a certification program, and expanded partner‑facing engineering and GTM resources to scale Claude’s enterprise adoption.

ByteDance reportedly pauses global launch of its Seedance 2.0 video generator. The company is delaying the planned mid‑March global rollout while engineers and lawyers add safeguards to address copyright complaints and cease‑and‑desist threats from studios.

Concerns

AI error jails innocent grandmother for months in North Dakota fraud case. She was wrongly identified by facial‑recognition software as a suspect in a Fargo bank fraud investigation, spent over five months jailed and extradited before her bank records proved she was 1,200 miles away and the charges were dropped.

Impossible to 100% prevent abuse, Grok lawyers say in Dutch case against nudify tools. Lawyers presented examples of Grok producing non‑consensual nude images and child sexual abuse material and asked the court to ban the feature with a €100,000‑per‑day fine, while xAI argued it cannot guarantee complete prevention of such misuse despite efforts to stop it.

Research

Many SWE-bench-Passing PRs Would Not Be Merged into Main. A maintainer‑reviewed evaluation of 296 SWE‑bench–passing AI‑generated PRs shows that about half would not be merged into main (after adjusting for maintainer noise), largely due to code‑quality issues, breaking other code, or core functionality problems that the automated grader missed.

Exclusive Self Attention. The proposed method, exclusive self attention (XSA), removes the component of attention outputs aligned with each token’s value vector to force attention to focus on contextual information, yielding consistent perplexity and downstream improvements with little extra cost across model sizes and sequence lengths.

Lost in Backpropagation: The LM Head is a Gradient Bottleneck. Findings show that the LM head’s low‑rank softmax causes a “gradient bottleneck” that compresses and destroys most of the backpropagated gradient—often 95–99%—slowing or preventing learning and reducing training efficiency by up to 16× even when model expressivity would otherwise suffice.

Thinking to Recall: How Reasoning Unlocks Parametric Knowledge in LLMs. Enabling chain‑of‑thought generation helps models surface related factual snippets and use the reasoning tokens as a computational buffer, which together increase parametric recall coverage but also introduce a hallucination risk that can be mitigated by selecting hallucination‑free reasoning trajectories at inference.

Omni-Diffusion: Unified Multimodal Understanding and Generation with Masked Discrete Diffusion. A joint mask‑based discrete diffusion over tokenized text, image, and speech enables any‑to‑any multimodal understanding and generation, with tailored training/inference techniques (progressive training, attenuated tail‑pad masking, position penalties, pre‑infilling, and adaptive token‑length initialization) that deliver comparable or better performance and faster sampling than existing autoregressive any‑to‑any systems on VQA, text‑to‑image, ASR, and TTS.

Neural Thickets: Diverse Task Experts Are Dense Around Pretrained Weights. The authors show that after pretraining, the parameter space near the weights becomes densely populated with diverse, task‑specialist perturbations, enabling simple random‑sampling plus ensembling (RandOpt) to achieve competitive post‑training improvements.

EndoCoT: Scaling Endogenous Chain-of-Thought Reasoning in Diffusion Models. This approach enables diffusion models to perform iterative, chain‑of‑thought‑style reasoning by updating MLLM latent states during generation and grounding the final reasoning state, improving accuracy on maze, TSP, VSP, and Sudoku tasks.

Our 236th episode with a summary and discussion of last week’s big AI news!

Recorded on 03/06/2026

Hosted by Andrey Kurenkov and Jeremie Harris

Feel free to email us your questions and feedback at andreyvkurenkov@gmail.com and/or hello@gladstone.ai

In this episode:

* OpenAI released GPT-5.4 Pro with a 1M-token context window, mid-response course correction, native computer-use capabilities, improved tool use, higher GPT-VAL performance (83%), and “high cyber capability” safety measures; OpenAI also launched GPT-5.3 Instant with a less “preachy” tone and a claimed 26.8% hallucination reduction.

* Google upgraded Gemini 3.1 Flash Lite with faster time-to-first-token and higher throughput, released a CLI for integrating agents with Gmail/Drive/Docs, and discussion highlighted real-world agent failure risks (including an example of an AI-driven mass email deletion).

* Luma launched unified multimodal models and Luma Agents for end-to-end creative work across text, image, video, and audio, including a reported ad localization use case completed in 40 hours for under $20,000.

* Defense-contract controversy escalated: Anthropic was labeled a supply chain risk (later narrowed), OpenAI’s DoD contract language emphasized “all lawful uses,” consumer cancellations boosted Claude’s app rankings, OpenAI saw departures and announced a $110B raise at a $730B valuation, Alibaba lost key Qwen leaders, a lawsuit alleged Gemini contributed to a suicide, Anthropic warned of major labor disruption, and METR corrected its AI time-horizon estimates.

A thank you to our current sponsors:

• Box - visit Box.com/AI to learn more

• ODSC AI - go to odsc.ai/east and use promo code LWAI for an additional 15% off your pass to ODSC AI East 2026.

• Factor - head to factormeals.com/lwai50off and use code lwai50off to get 50 percent off and free breakfast for a year

PS my company Astrocade is hiring for engineers, marketing, product, growth, and more! If you’re in the bay area, would like to join a small but growing startup, and think building a youtube-of-games sounds exciting, feel free to email me at andrey@astroblox.ai or message me on LinkedIn.

Timestamps:

• (00:00:10) Intro / Banter

• (00:01:19) News Preview

• Tools & Apps

• (00:02:10) OpenAI launches GPT-5.4 with Pro and Thinking versions | TechCrunch

• (00:12:31) OpenAI GPT-5.3 Instant less likely to beat around the bush • The Register

• (00:16:07) Google releases Gemini 3.1 Flash Lite at 1/8th the cost of Pro | VentureBeat

• (00:19:23) Google makes Gmail, Drive, and Docs ‘agent-ready’ for OpenClaw | PCWorld

• (00:27:02) Luma launches creative AI agents powered by its new ‘Unified Intelligence’ models | TechCrunch

• Applications & Business

• (00:30:05) Anthropic CEO Dario Amodei calls OpenAI’s messaging around military deal ‘straight up lies,’ report says | TechCrunch

• (00:41:56) No ethics at all’: the ‘cancel ChatGPT’ trend is growing after OpenAI signs a deal with the US military | TechRadar

• (00:45:54) OpenAI raises $110B in one of the largest private funding rounds in history | TechCrunch

• (00:56:07) Alibaba scrambles after sudden departure of Qwen tech lead

• Policy & Safety

• (01:00:12) Pentagon approves OpenAI safety red lines after dumping Anthropic + Where things stand with the Department of War Anthropic + Microsoft says Anthropic’s products remain available to customers after Pentagon blacklist

• (01:09:11) A new lawsuit claims Gemini assisted in suicide | Semafor

• (01:15:24) Anthropic just mapped out which jobs AI could potentially replace. A ‘Great Recession for white-collar workers’ is absolutely possible | Fortune

• (01:21:54) We’re correcting a mistake in our modeling that inflated recent 50%-time horizons by 10-20%

PS we are hiring for engineers, marketing, product, growth, and more! If you’re in the bay area, would like to join a small but growing startup, and think building a youtube-of-games sounds exciting, feel free to email me at andrey@astroblox.ai or message me on LinkedIn.

Top News

Anthropic officially told by DOD that it’s a supply chain risk even as Claude used in Iran

Related:

• Anthropic Hits Back After US Military Labels It a ‘Supply Chain Risk’

• Anthropic CEO: We’re trying to “deescalate” Pentagon AI standoff to reach “some agreement that works for us and works for them”

• Where things stand with the Department of War

• Anthropic CEO Dario Amodei calls OpenAI’s messaging around military deal ‘straight up lies,’ report says

• ‘No ethics at all’: the ‘cancel ChatGPT’ trend is growing after OpenAI signs a deal with the US military

• Microsoft, Google, Amazon say Anthropic Claude remains available to non-defense customers

Summary: The Pentagon formally designated Anthropic a “supply chain risk,” ordering defense vendors and contractors to certify they are not using Claude in DoD work and triggering an immediate wind-down of Pentagon use of Anthropic’s models. The dispute centers on contract terms: Anthropic sought explicit red lines against mass domestic surveillance and fully autonomous weapons, while the DoD insisted on access for “all lawful purposes.” Anthropic says the designation under 10 USC 3252 is narrow, applies only to uses tied directly to DoD contracts, and violates the statute’s “least restrictive means” requirement; the company plans to challenge it in court. At the same time, multiple reports indicate Claude has supported U.S. military operations in Iran for intelligence analysis, modeling and simulation, operational planning, and cyber operations.

After Defense Secretary Pete Hegseth suggested no military partner may conduct any commercial activity with Anthropic, confusion spread about the scope. Cloud providers disputed that interpretation; Microsoft, Google, and AWS clarified Claude remains available for non-defense workloads via M365, GitHub, AI Foundry, and Google Cloud.

Meanwhile, OpenAI announced a DoD deal to deploy its models in classified environments, citing prohibitions on mass surveillance and autonomous weapon systems. Anthropic CEO Dario Amodei called OpenAI’s framing “straight up lies” and “safety theater,” arguing the agreement still hinges on “all lawful purposes.”

Editor’s Take: The designation of Anthropic as a supply chain risk could have meant a massive hit to its business, but as of now it appears that may not happen. In fact, Anthropic might on the whole have benefited from the events, given the massive rise in consumer awareness of Claude as an alternative to ChatGPT. Still, it appears likely that this story is not yet over.

‘No ethics at all’: the ‘cancel ChatGPT’ trend is growing after OpenAI signs a deal with the US military

Related:

• Anthropic’s Claude rises to No. 1 in the App Store following Pentagon dispute

• How OpenAI caved to the Pentagon on AI surveillance

• ‘QuitGPT’ protesters rally outside OpenAI HQ in San Francisco over deal with Pentagon

• OpenAI alters deal with Pentagon as critics sound alarm over surveillance

• OpenAI hardware exec Caitlin Kalinowski quits in response to Pentagon deal

Summary: OpenAI signed an agreement with the Department of War after Anthropic walked away over, as covered above. OpenAI says its deal enforces “red lines” via legal references and technical guardrails, citing compliance with the Fourth Amendment, FISA, EO 12333, and DoD policy, plus cloud-only deployment, security clearances, and classifiers to monitor use.

Critics counter that the terms still hinge on “any lawful use,” a standard that has historically enabled bulk data programs like PRISM and Verizon call record collection, and they note vague carve-outs such as prohibitions on “unconstrained” or “generalized” monitoring. OpenAI later posted revised language that “the AI system shall not be intentionally used for domestic surveillance of U.S. persons,” and said defense intelligence components (e.g., NSA) are excluded, but it has not released the full contract; legal experts warn loopholes remain and that policy reinterpretations can broaden surveillance.

The fallout was swift: a “QuitGPT” boycott, in-person protests at OpenAI’s SF HQ with “No killer robots” messaging, and a 295% surge in ChatGPT uninstalls. Anthropic’s Claude climbed to No. 1 in the U.S. App Store, with daily sign-ups breaking records, free users up 60% since January, and paid subscribers more than doubling this year. Inside OpenAI, hardware lead Caitlin Kalinowski resigned, citing rushed governance and insufficiently defined guardrails on surveillance and lethal autonomy.

Editor’s Take: Consumer boycotts like “QuitGPT” rarely lead to a significant lasting loss for the targeted company, but Claude has thus far been the underdog relative to ChatGPT with consumers, and the scale of the reaction does appear meaningful. In either case, the impact on Anthropic’s hiring power is likely even more significant given that how hotly fought for top tier AI talent is.

OpenAI launches GPT-5.4 with Pro and Thinking versions

Related:

• ‘We heard your feedback loud and clear’ — OpenAI introduces new ChatGPT 5.3 Instant to ‘reduce the cringe’ for all users

Summary: OpenAI released GPT-5.4 in three variants—standard, Pro, and Thinking—positioning it as a more capable and efficient frontier model for professional work. The API now supports up to a 1 million-token context window, and improved token efficiency means it can solve the same tasks with fewer tokens than prior models. GPT-5.4 posted record results on OSWorld-Verified and WebArena Verified for computer use, scored 83% on OpenAI’s GDPval for knowledge work, and led Mercor’s APEX-Agents benchmark for law and finance tasks. OpenAI also introduced Tool Search for API tool calling, allowing the model to fetch tool definitions on demand to reduce token overhead and latency, and reported fewer factual errors versus GPT-5.2 (33% fewer erroneous claims; 18% fewer responses with errors).

Separately, OpenAI upgraded ChatGPT’s default to GPT-5.3 Instant to “reduce the cringe,” tuning away unnecessary refusals and long safety disclaimers while keeping guardrails. GPT-5.3 Instant focuses on direct answers for reasonable requests and improves accuracy, with ~27% fewer hallucinations during online research and ~20% fewer without browsing, plus better blending of web data with prior knowledge. The change updates the baseline chat experience across common uses while retaining GPT-5.2 Instant for paid users as a legacy option.

Editor’ Take: I discussed the accelerating rate of these ‘+0.1’ model releases in the last couple of these newsletters, and the same comments largely apply here. Relative to last year, 2026

Other News

Tools

Luma launches creative AI agents powered by its new ‘Unified Intelligence’ models. The agents can plan, generate, and iteratively refine text, images, video, and audio for ad agencies and brands using a single multimodal Uni‑1 model that maintains persistent context across assets and coordinates with other AI models.

Anthropic’s Claude reports widespread outage. Thousands of users reported login and access failures affecting Claude.ai and Claude Code, while Anthropic said the API remained operational and engineers were implementing a fix.

Google makes Gmail, Drive, and Docs ‘agent-ready’ for OpenClaw. A new CLI offers developer‑focused tools and documentation to simplify and standardize how AI agents like OpenClaw and MCP‑compatible apps connect to Gmail, Drive, Docs, and other Workspace services.

Google reveals dev-focused Gemini 3.1 Flash Lite, promises ‘best-in-class intelligence for your highest-volume workloads’. The model offers faster response and output generation while cutting per‑token input/output costs versus Gemini 2.5 Flash, outperforms several rivals on multiple benchmarks, and is available to developers in preview via the Gemini API and Vertex AI.

Cursor is rolling out a new kind of agentic coding tool. The Automations feature lets engineers trigger and manage always‑on agents from events like code changes, Slack messages, timers, PagerDuty incidents, and weekly summaries so agents can run reviews, security audits, incident responses, and other recurring tasks without constant human initiation.

Business

Alibaba scrambles after sudden departure of Qwen tech lead. The sudden departure sparked urgent all‑hands discussions about restructuring, resource allocation, and whether new hires like Hao Zhou will fill Lin’s technical and leadership roles.

OpenAI raises $110B in one of the largest private funding rounds in history. The funding includes large compute and services commitments—notably AWS and Nvidia infrastructure partnerships that pledge gigawatts of training/inference capacity and substantial Bedrock integration—while the round remains open and some payments are contingent on future conditions like AGI or an IPO.

Cursor has reportedly surpassed $2B in annualized revenue. Enterprise customers now account for about 60% of revenue, helping offset individual‑user defections and driving a recent doubling in the company’s revenue run rate.

Research

A Very Big Video Reasoning Suite. VBVR‑Dataset (1,007,500 clips) spans 200 tasks across five cognitive pillars, paired with a verifiable VBVR‑Bench aligned with human judgments; a scaling study shows improved yet still sub‑human video reasoning as model and data scale increase.

Exploratory Memory-Augmented LLM Agent via Hybrid On- and Off-Policy Optimization. The method jointly trains model parameters with on‑ and off‑policy RL while updating a non‑parametric memory during interaction, using memory‑guided and memory‑free rollouts to drive broader exploration and improve performance on embodied reasoning benchmarks.

HyperSteer: Activation Steering at Scale with Hypernetworks. Hypernetworks generate task‑conditioned steering vectors for an instruction‑tuned base LM, achieving better and more scalable activation steering than prior methods and matching steering‑via‑prompting on held‑out prompts.

Agentic Code Reasoning. Prompting LLM agents with semi‑formal structured certificates—explicit premises, traced code paths, and per‑test evidence—improves execution‑free patch equivalence verification, code QA, and fault localization across diverse benchmarks without formal semantics or specialized training.

Beyond Language Modeling: An Exploration of Multimodal Pretraining. Training a single model from scratch with mixed text, image, and video data—using semantic representation autoencoders plus MoE—shows joint vision‑language learning without fatal interference, reveals a vision data hunger and scaling asymmetry mitigated by sparsity, and demonstrates that video‑rich multimodal pretraining yields emergent world‑modeling and generation/understanding abilities.

Semantic Tube Prediction: Beating LLM Data Efficiency with JEPA. Adding an STP auxiliary loss that minimizes components of hidden‑state deviation perpendicular to inferred geodesic trajectories (“semantic tube”) improves signal‑to‑noise ratio, data efficiency, and diversity in autoregressive LLM training compared to next‑token prediction alone.

The Spike, the Sparse and the Sink: Anatomy of Massive Activations and Attention Sinks. Massive outlier activations and attention sinks commonly co‑occur in pretrained decoder‑only Transformers due to specific normalization choices, attention‑space dimensionality, and context‑length training; each phenomenon can be independently mitigated without hurting language‑modeling performance.

Progressive Residual Warmup for Language Model Pretraining. ProRes gradually scales each layer’s residual from zero to one during training—delaying deeper layers’ contributions so shallow layers stabilize first—improving perplexity and depth scaling across models from 71M to 7B parameters.

Reasoning Models Struggle to Control their Chains of Thought. CoT‑Control, a 14,076‑instance benchmark, evaluates how well models follow explicit instructions to modify their chains of thought; most reasoning models struggle with controllability, which varies with model size, reasoning effort, training, and situational‑awareness prompting.

Analysis

Anthropic reports on trends in occupations susceptible to AI and whether there are changes in the labor market. Anthropic’s analysis introduces an “observed exposure” metric showing which occupations (notably programmers, customer service reps, and financial analysts) have tasks both theoretically automatable by AI and already automated in practice, finds higher exposure among better‑educated, higher‑income workers, and reports little current impact on unemployment overall but a small decline in employment among 22–25‑year‑olds in high‑exposure roles.

Our 235th episode with a summary and discussion of last week’s big AI news!

Recorded on 02/27/2026. Hosted by Andrey Kurenkov and Jeremie Harris

Note from Andrey: my startup Astrocade is hiring for engineers, marketing, product, growth, and more! If you’re in the bay area, would like to join a small but growing startup, and think building a youtube-of-games sounds exciting, feel free to email me at andrey@astroblox.ai

Feel free to email us your questions and feedback at andreyvkurenkov@gmail.com and/or hello@gladstone.ai

In this episode:

• Model and tool updates highlight Anthropic’s Sonnet 4.6 (1M context; strong ARC-AGI-2 results), Google’s Gemini 3.1 Pro (major ARC-AGI-2 jump and multimodal demos), xAI’s Grok 4.2 beta (multi-agent debate), plus Anthropic’s Claude Code “Remote Control” and Perplexity’s multi-agent “Computer” coordinator.

• Compute and business moves include Meta’s reported up-to-$100B AMD chip deal with warrant/equity incentives, MatX raising $500M to build specialized transformer chips shipping in 2027, World Labs raising $1B for world-model/3D environment tech, and a new startup raising $100M to simulate/predict human behavior.

• Infrastructure and geopolitics cover Stargate data-center delays amid OpenAI/Oracle/SoftBank control disputes and cash concerns, and China’s plan to scale 7nm/5nm wafer output despite yield and tooling constraints.

• Research and safety/policy discuss optimizer gains from masked updates, “deep thinking tokens” as a reasoning-effort signal, LLM attractor-state behaviors in bot-to-bot chats, mechanistic interpretability of counting/line-wrapping, methods to map task difficulty to human time horizons, plus Anthropic–Pentagon contract tensions, Anthropic’s report on distillation attacks (DeepSeek/Moonshot/Minimax), and OpenAI’s report on disrupting malicious use.

A thank you to our current sponsors:

• Box - visit Box.com/AI to learn more

• ODSC AI - go to odsc.ai/east and use promo code LWAI for an additional 15% off your pass to ODSC AI East 2026.

• Factor - head to factormeals.com/lwai50off and use code lwai50off to get 50 percent off and free breakfast for a year

Timestamps:

• (00:00:10) Intro / Banter

• (00:01:52) News Preview
  

  Tools & Apps

• (00:03:20) Anthropic releases Sonnet 4.6 | TechCrunch

• (00:11:24) Google Rolls Out Latest AI Model, Gemini 3.1 Pro - CNET

• (00:14:54) Elon Musk says Grok 4.20 public beta is now available: Capabilities of AI chatbot offered by xAI - The Times of India

• (00:18:06) Anthropic just released a mobile version of Claude Code called Remote Control | VentureBeat

• (00:21:01) Perplexity announces “Computer,” an AI agent that assigns work to other AI agents - Ars Technica
  

  Applications & Business

• (00:23:40) Meta strikes up to $100B AMD chip deal as it chases ‘personal superintelligence’ | TechCrunch

• (00:27:05) Nvidia challenger AI chip startup MatX raised $500M | TechCrunch

• (00:31:00) World Labs lands $1B, with $200M from Autodesk, to bring world models into 3D workflows | TechCrunch

• (00:33:07) Simile Raises $100 Million for AI Aiming to Predict Human Behavior

• (00:33:52) Stargate AI data centers for OpenAI reportedly delayed by squabbles between partners — sources say OpenAI, Oracle, and SoftBank disagreed on who would have ultimate control of the planned data centers

• (00:36:43) China to increase leading-edge chip output by 5x in two years, report claims — aims to lift 7nm and 5nm production to 100,000 wafers per month, targeting half a million monthly by 2030
  

  Research & Advancements

• (00:40:33) On Surprising Effectiveness of Masking Updates in Adaptive Optimizers

• (00:48:03) Think Deep, Not Just Long: Measuring LLM Reasoning Effort via Deep-Thinking Tokens

• (00:54:52) models have some pretty funny attractor states

• (01:01:41) When Models Manipulate Manifolds: The Geometry of a Counting Task

• (01:05:16) BRIDGE: Predicting Human Task Completion Time From Model Performance

• (01:12:00) NESSiE: The Necessary Safety Benchmark -- Identifying Errors that should not Exist

• (01:13:15) The least understood driver of AI progress

• (01:21:45) The Persona Selection Model: Why AI Assistants might Behave like Humans
  

  Policy & Safety

• (01:25:04) Anthropic CEO Amodei says Pentagon’s threats ‘do not change our position’ on AI

• (01:33:04) Musk’s xAI, Pentagon reach deal to use Grok in classified systems

• (01:34:17) Detecting and preventing distillation attacks

• (01:38:36) OpenAI details expanding efforts to disrupt malicious use of AI in new report - SiliconANGLE

Related:

• Claude Sonnet 4.6 model brings ‘much-improved coding skills’ and upgraded free tier

• Claude Sonnet 4.6 delivers frontier-level AI for free and cheap-seat users

• Anthropic releases Claude Sonnet 4.6, continuing breakneck pace of AI model releases

Read more

Our 234th episode with a summary and discussion of last week’s big AI news!

Recorded on 01/02/2026

Hosted by Andrey Kurenkov and Jeremie Harris

Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai

In this episode:

• Major model launches include Anthropic’s Opus 4.6 with a 1M-token context window and “agent teams,” OpenAI’s GPT-5.3 Codex and faster Codex Spark via Cerebras, and Google’s Gemini 3 Deep Think posting big jumps on ARC-AGI-2 and other STEM benchmarks amid criticism about missing safety documentation.

• Generative media advances feature ByteDance’s Seedance 2.0 text-to-video with high realism and broad prompting inputs, new image models Seedream 5.0 and Alibaba’s Qwen Image 2.0, plus xAI’s Grok Imagine API for text/image-to-video.

• Open and competitive releases expand with Zhipu’s GLM-5, DeepSeek’s 1M-token context model, Cursor Composer 1.5, and open-weight Qwen3 Coder Next using hybrid attention aimed at efficient local/agentic coding.

• Business updates include ElevenLabs raising $500M at an $11B valuation, Runway raising $315M at a $5.3B valuation, humanoid robotics firm Apptronik raising $935M at a $5.3B valuation, Waymo announcing readiness for high-volume production of its 6th-gen hardware, plus industry drama around Anthropic’s Super Bowl ad and departures from xAI.

Timestamps:

• (00:00:10) Intro / Banter

• (00:02:05) Response to listener comments

Tools & Apps

• (00:03:59) Anthropic releases Opus 4.6 with new ‘agent teams’ | TechCrunch

• (00:08:00) OpenAI’s new GPT-5.3-Codex is 25% faster and goes way beyond coding now - what’s new | ZDNET

• (00:22:02) OpenAI launches new macOS app for agentic coding | TechCrunch

• (00:23:10) Google Unveils Gemini 3 Deep Think for Science & Engineering | The Tech Buzz

• (00:27:58) ByteDance’s Seedance 2.0 Might be the Best AI Video Generator Yet - TechEBlog

• (00:31:46) China’s ByteDance, Alibaba unveil AI image tools to rival Google’s popular Nano Banana | South China Morning Post

• (00:33:26) DeepSeek boosts AI model with 10-fold token addition as Zhipu AI unveils GLM-5 | South China Morning Post

• (00:39:43) Cursor launches Composer 1.5 with upgrades for complex tasks

• (00:40:35) xAI launches Grok Imagine API for text and image to video

Applications & Business

• (00:42:19) Nvidia-backed AI voice startups ElevenLabs hits $11 billion valuation

• (00:48:36) AI video startup Runway raises $315M at $5.3B valuation, eyes more capable world models | TechCrunch

• (00:50:34) Humanoid robot startup Apptronik has now raised $935M at a $5B+ valuation | TechCrunch

• (00:53:42) Anthropic says ‘Claude will remain ad-free,’ unlike an unnamed rival | The Verge

• (00:56:50) Okay, now exactly half of xAI’s founding team has left the company | TechCrunch

• (01:00:35) Waymo’s next-gen robotaxi is ready for passengers — and also ‘high-volume production’ | The Verge

Projects & Open Source

• (01:01:31) Qwen3-Coder-Next: Pushing Small Hybrid Models on Agentic Coding

• (01:05:10) OpenClaw’s AI ‘skill’ extensions are a security nightmare | The Verge

Research & Advancements

• (01:07:12) Learning to Reason in 13 Parameters

• (01:12:33) Reinforcement World Model Learning for LLM-based Agents

• (01:16:32) Opus 4.6 on Vending-Bench – Not Just a Helpful Assistant

Policy & Safety

• (01:19:00) METR GPT-5.2

(01:23:31) The Hot Mess of AI: How Does Misalignment Scale with Model Intelligence and Task Complexity?

I am also going to be adding an ‘Editor’s Take’ for Top News to add a bit commentary and extra cont…

Read more

Our 233rd episode with a summary and discussion of last week’s big AI news!

Recorded on 01/30/2026

Hosted by Andrey Kurenkov and Jeremie Harris

Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai

In this episode:

• Google introduces Gemini AI agent in Chrome for advanced browser functionality, including auto-browsing for pro and ultra subscribers.

• OpenAI releases ChatGPT Translator and Prism, expanding its applications beyond core business to language translation and scientific research assistance.

• Significant funding rounds and valuations achieved by startups Recursive and New Rofo, focusing on specialized AI chips and optical processors respectively.

• Political and social issues, including violence in Minnesota, prompt tech leaders in AI like Ade from Anthropic and Jeff Dean from Google to express concerns about the current administration’s actions.

Timestamps:

• (00:00:10) Intro / Banter

• Tools & Apps

• (00:04:09) Google adds Gemini AI-powered ‘auto browse’ to Chrome | The Verge

• (00:07:11) Users flock to open source Moltbot for always-on AI, despite major risks - Ars Technica

• (00:13:25) Google Brings Genie 3 ‘World Building’ Experiment to AI Ultra Subscribers - CNET

• (00:16:17) OpenAI’s ChatGPT translator challenges Google Translate | The Verge

• (00:18:27) OpenAI launches Prism, a new AI workspace for scientists | TechCrunch

• Applications & Business

• (00:19:49) Exclusive: China gives nod to ByteDance, Alibaba and Tencent to buy Nvidia’s H200 chips - sources | Reuters

• (00:22:55) AI chip startup Ricursive hits $4B valuation 2 months after launch

• (00:24:38) AI Startup Recursive in Funding Talks at $4 Billion Valuation - Bloomberg

• (00:27:30) Flapping Airplanes and the promise of research-driven AI | TechCrunch

• (00:31:54) From invisibility cloaks to AI chips: Neurophos raises $110M to build tiny optical processors for inferencing | TechCrunch

• Projects & Open Source

• (00:35:34) Qwen3-Max-Thinking debuts with focus on hard math, code

• (00:38:26) China’s Moonshot releases a new open-source model Kimi K2.5 and a coding agent | TechCrunch

• (00:46:00) Ai2 launches family of open-source AI developer agents that adapt to any codebase - SiliconANGLE

• (00:47:46) Tiny startup Arcee AI built a 400B-parameter open source LLM from scratch to best Meta’s Llama

• Research & Advancements

• (00:52:53) Post-LayerNorm Is Back: Stable, ExpressivE, and Deep

• (00:58:00) [2601.19897] Self-Distillation Enables Continual Learning

• (01:03:04) [2601.20802] Reinforcement Learning via Self-Distillation

• (01:05:58) Teaching Models to Teach Themselves: Reasoning at the Edge of Learnability

• Policy & Safety

• (01:09:13) Amodei, Hoffman Join Tech Workers Decrying Minnesota Violence - Bloomberg

Moonshot AI unveiled Kimi K2.5, an open-source, natively multimodal model trained on 15 trillion mixed visual and text tokens that understands text, images, and video. The company emphasizes strong agentic capabilities, citing “agent swarm” orchestration where multiple agents …

Read more

Our 232st episode with a summary and discussion of last week’s big AI news!

Recorded on 01/23/2026

Hosted by Andrey Kurenkov and Jeremie Harris

Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai

In this episode:

• OpenAI announces testing of ads in ChatGPT and introduces child age prediction to enhance safety features, amidst ongoing ethical debates and funding expansions in AI integration with educational tools and business models.

• China’s AI landscape sees significant progress with AI firm Jpu training advanced models on domestic hardware, and strong competitive moves by data centers, highlighting the intense demand in AI manufacturing and infrastructure.

• Silicon Valley tensions rise as startup Thinking Machines experiences high-profile departures back to OpenAI, reflecting broader industry struggles and rapid shifts in organizational dynamics.

• AI legislation and safety measures advance with the US Senate’s Defiance Act addressing explicit content, and Anthropic updating Claude’s constitution to guide ethical AI interactions, while cultural pushbacks from artists signal ongoing debates in intellectual property and AI-generated content.

Timestamps:

• (00:00:10) Intro / Banter

• (00:02:08) News Preview

• (00:02:26) Response to listener comments

• Tools & Apps

• (00:11:55) OpenAI to test ads in ChatGPT as it burns through billions - Ars Technica

• (00:18:05) OpenAI is launching age prediction for ChatGPT accounts

• (00:23:37) Google now offers free SAT practice exams, powered by Gemini | TechCrunch

• (00:24:57) Baidu’s AI Assistant Reaches Milestone of 200 Million Monthly Active Users - WSJ

• Applications & Business

• (00:26:53) The Drama at Thinking Machines, a New A.I. Start-Up, Is Riveting Silicon Valley - The New York Times

• (00:31:44) Zhipu AI breaks US chip reliance with first major model trained on Huawei stack | South China Morning Post

• (00:36:31) Elon Musk’s xAI launches world’s first Gigawatt AI supercluster to rival OpenAI and Anthropic

• (00:41:25) Sequoia to invest in Anthropic, breaking VC taboo on backing rivals: FT

• (00:45:18) Humans&, a ‘human-centric’ AI startup founded by Anthropic, xAI, Google alums, raised $480M seed round | TechCrunch

• Projects & Open Source

• (00:48:51) Black Forest Labs Releases FLUX.2 [klein]: Compact Flow Models for Interactive Visual Intelligence - MarkTechPost

• (00:50:35) [2601.10611] Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding

• (00:52:53) [2601.10547] HeartMuLa: A Family of Open Sourced Music Foundation Models

• (00:54:46) [2601.11044] AgencyBench: Benchmarking the Frontiers of Autonomous Agents in 1M-Token Real-World Contexts

• Research & Advancements

• (00:57:05) STEM: Scaling Transformers with Embedding Modules

• (01:06:22) Reasoning Models Generate Societies of Thought

• (01:14:21) Why LLMs Aren’t Scientists Yet: Lessons from Four Autonomous Research Attempts

• Policy & Safety

• (01:19:41) Senate passes bill letting victims sue over Grok AI explicit images

• (01:22:03) Building Production-Ready Probes For Gemini

• (01:27:32) Anthropic Publishes Claude AI’s New Constitution | TIME

• Synthetic Media & Art

• (01:34:13) Artists Launch Stealing Isn’t Innovation Campaign To Protest Big Tech

Related:

• ChatGPT to begin testing ads as generative AI competition heats up

OpenAI will begin testing labeled banner ads in ChatGPT for logged‑in users on the free tier and the $8/month ChatGPT Go plan, rolling out in the U.S. and other markets in the coming weeks. Ads will appear as blocked-off se…

Read more

Our 231st episode with a summary and discussion of last week’s big AI news!

Recorded on 01/16/2026

Hosted by Andrey Kurenkov and Jeremie Harris

Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai

In this episode:

• Anthropic’s new cowork tool integrates Claude code, potentially simplifying multiple computing tasks from editing videos to compiling spreadsheets.

• Significant funding rounds see Anthropic raising $10B at a valuation of $350B, while XAI raises $20B, underscoring the immense market interest in AI startups.

• Nvidia faces supply challenges for H200 AI chips due to overwhelming demand from China, despite high costs per unit and its potential impact on U.S. company revenue.

• Policy debates highlight tensions around U.S. export controls to China, with leaders like Justin Lin from Alibaba and Jake Sullivan, former national security advisor, weighing in on the ramifications for the AI industry’s future.

Timestamps:

• (00:00:10) Intro / Banter

• (00:01:30) News Preview

• Tools & Apps

• (00:02:13) Anthropic’s new Cowork tool offers Claude Code without the code | TechCrunch

• (00:09:45) Google’s Gemini AI will use what it knows about you from Gmail, Search, and YouTube | The Verge

• (00:12:45) Google removes some AI health summaries after investigation finds “dangerous” flaws - Ars Technica

• (00:16:29) Gmail is getting a Gemini AI overhaul

• (00:18:12) Slackbot is an AI agent now | TechCrunch

• Applications & Business

• (00:20:11) Anthropic Raising $10 Billion at $350 Billion Value

• (00:22:25) Elon Musk xAI raises $20 billion from Nvidia, Cisco, investors

• (00:24:47) NVIDIA Needs a Supply Chain ‘Miracle’ From TSMC as China’s H200 AI Chip Orders Overwhelm Supply, Triggering a Bottleneck

• (00:29:26) OpenAI signs deal, worth $10B, for compute from Cerebras | TechCrunch

• (00:31:49) CoreWeave in focus as it amends credit agreement

• (00:34:30) LMArena lands $1.7B valuation four months after launching its product | TechCrunch

• Projects & Open Source

• (00:35:54) Nemotron-Cascade: Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Models

• (00:43:15) mHC: Manifold-Constrained Hyper-Connections

• (00:49:53) IQuest_Coder_Technical_Report

• (00:54:58) TII Abu-Dhabi Released Falcon H1R-7B: A New Reasoning Model Outperforming Others in Math and Coding with only 7B Params with 256k Context Window - MarkTechPost

• Research & Advancements

• (01:01:42) Deep Delta Learning

• (01:07:47) Recursive Language Models

• (01:13:39) Conditional memory via scalable lookup

• (01:18:54) Extending the Context of Pretrained LLMs by Dropping their Positional Embeddings

• Policy & Safety

• (01:26:06) Constitutional Classifiers++: Efficient Production-Grade Defenses against Universal Jailbreaks

• (01:31:00) Nvidia CEO says purchase orders, not formal declaration, will signal Chinese approval of H200

• (01:32:24) China AI Leaders Warn of Widening Gap With US After $1B IPO Week

• (01:37:25) Jake Sullivan is furious that Trump removed Biden’s AI chip export controls | The Verge

Apple announced a multi-year partnership to use Google’s Gemini models and Google Cloud to power AI features like Siri, after testing alternatives from OpenAI and Anthropic. According to both companies, Gemini provides “the most capable foundation” for Apple’s own models, with reporting suggesting Ap…

Read more

Our 230th episode with a summary and discussion of last week’s big AI news!

Recorded on 01/02/2026

Hosted by Andrey Kurenkov and Jeremie Harris

Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai

In this episode:

• Nvidia’s acquisition of AI chip startup Groq for $20 billion highlights a strategic move for enhanced inference technology in GPUs.

• New York’s RAISE Act legislation aims to regulate AI safety, marking the second major AI safety bill in the US.

• The launch of GLM 4.7 by Zhipu AI marks a significant advancement in open-source AI models for coding.

• Evaluation of long-horizon AI agents raises concerns about the rising costs and efficiency of AI in performing extended tasks.

Timestamps:

• (00:00:10) Intro / Banter

• (00:01:58) 2025 Retrospective

• Tools & Apps

  • (00:24:39) OpenAI bets big on audio as Silicon Valley declares war on screens | TechCrunch

• Applications & Business

  • (00:26:39) Nvidia buying AI chip startup Groq for about $20 billion, biggest deal

  • (00:34:28) Exclusive | Meta Buys AI Startup Manus, Adding Millions of Paying Users - WSJ

  • (00:38:05) Cursor continues acquisition spree with Graphite deal | TechCrunch

  • (00:39:15) Micron Hikes CapEx to $20B with 2026 HBM Supply Fully Booked; HBM4 Ramps 2Q26

  • (00:42:06) Chinese fabs are reportedly upgrading older ASML DUV lithography chipmaking machines — secondary channels and independent engineers used to soup up Twinscan NXT series

• Projects & Open Source

  • (00:47:52) Z.AI launches GLM-4.7, new SOTA open-source model for coding

  • (00:50:11) Evaluating AI’s ability to perform scientific research tasks

• Research & Advancements

  • (00:54:32) Large Causal Models from Large Language Models

  • (00:57:33) Universally Converging Representations of Matter Across Scientific Foundation Models

  • (01:02:11) META-RL INDUCES EXPLORATION IN LANGUAGE AGENTS

  • (01:07:16) Are the Costs of AI Agents Also Rising Exponentially?

  • (01:11:17) METR eval for Opus 4.5

  • (01:16:19) How to game the METR plot

• Policy & Safety

  • (01:17:24) New York governor Kathy Hochul signs RAISE Act to regulate AI safety | TechCrunch

  • (01:20:40) Activation Oracles: Training and Evaluating LLMs as General-Purpose Activation Explainers

  • (01:26:46) Monitoring Monitorability

  • (01:32:07) Sam Altman is hiring someone to worry about the dangers of AI | The Verge

  • (01:33:38) X users asking Grok to put this girl in bikini, Grok is happy obliging - India Today

Related:

• Nvidia launches Alpamayo, open AI models that allow autonomous vehicles to ‘think like a human’

• Nvidia launches Vera Rubin AI computing platform at CES 2026

At CES 2026, Nvidia CEO Jensen Huang announced the company’s new AI chip, Vera Rubin, which will begin shipping to custome…

Read more

Our 229th episode with a summary and discussion of last week’s big AI news!

Recorded on 12/19/2025

Hosted by Andrey Kurenkov and Jeremie Harris

Feel free to email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai

In this episode:

• Notable releases include OpenAI’s GPT-5.2 Codex for advanced coding and Google’s Gemini Free Flash for competitive AI application performance. Nvidia’s new open-source Trion-3 models also showcase impressive benchmarks.

• Funding updates highlight Lovable’s $330M Series B, valuing the AI coding startup at $6.6B, and Faya’s $140M Series D for AI model hosting, valued at $4.5B.

• China makes significant strides in semiconductor technology with advances in EUV lithography machines, led by Huawei and SMIC, potentially disrupting global chip manufacturing dominance.

• Key safety and policy updates include OpenAI’s GPT-5.2 system card focusing on biosecurity and cybersecurity risks, while Google partners with the US military to power a new AI platform with Gemini models.

Timestamps:

• (00:00:10) Intro / Banter

• (00:02:09) News Preview
  

• Tools & Apps

• (00:02:56) Google launches Gemini 3 Flash, makes it the default model in the Gemini app | TechCrunch

• (00:10:13) ChatGPT launches an app store, lets developers know it’s open for business | TechCrunch

• (00:13:35) Introducing GPT-5.2-Codex | OpenAI

• (00:19:23) Story about OpenAI release - GPT image 1.5

• (00:22:27) Meta partners with ElevenLabs to power AI audio across Instagram, Horizon - The Economic Times
  

• Applications & Business

• (00:23:16) OpenAI to End Equity Vesting Period for Employees, WSJ Says

• (00:28:20) How China built its ‘Manhattan Project’ to rival the West in AI chips

• (00:36:47) China’s Huawei, SMIC Make Progress With Chips, Report Finds

• (00:41:03) OpenAI in Talks to Raise At Least $10 Billion From Amazon and Use Its AI Chips

• (00:43:32) Amazon has a new leader for its ‘AGI’ group as it plays catch-up on AI | The Verge

• (00:47:27) Broadcom reveals its mystery $10 billion customer is Anthropic

• (00:49:12) Vibe-coding startup Lovable raises $330M at a $6.6B valuation | TechCrunch

• (00:50:38) Fal nabs $140M in fresh funding led by Sequoia, tripling valuation to $4.5B | TechCrunch
  

• Projects & Open Source

• (00:51:10) Nvidia Becomes a Major Model Maker With Nemotron 3 | WIRED

• (00:59:24) Meta introduces new SAM AI able to isolate and edit audio • The Register

• (00:59:54) [2512.14856] T5Gemma 2: Seeing, Reading, and Understanding Longer

• (01:03:10) Anthropic makes agent Skills an open standard - SiliconANGLE
  

• Research & Advancements

• (01:03:47) Budget-Aware Tool-Use Enables Effective Agent Scaling

• (01:08:21) Rethinking Thinking Tokens: LLMs as Improvement Operators

• (01:10:50) What if AI capabilities suddenly accelerated in 2027? How would the world know?
  

• Policy & Safety

• (01:12:58) Update to GPdfT-5 System Card: GPT-5.2

• (01:18:04) Neural Chameleons: Language Models Can Learn to Hide Their Thoughts from Unseen Activation Monitors

• (01:20:47) Async Control: Stress-testing Asynchronous Control Measures for LLM Agents

• (01:24:37) Google is powering a new US military AI platform | The Verge

Nvidia agreed to license Groq’s inference technology and acquire essentially all of its assets for about $20 billion in cash—its largest deal to date—according to Disruptive CEO Alex Davis, a major Groq investor. Groq, valued at $6.9 billion after raising $750 mil…

Read more