Subvolts take Track AI's rapid model and hardware advancements, from Anthropic's Sonnet to major chip deals. Mar 3, 2026
For you
Developers
Learn about new model capabilities like larger context windows and tools for coding and agentic tasks.
Business
Understand major AI hardware investments, strategic partnerships, and the evolving competitive landscape.
Everyday users
Stay informed on the latest AI model improvements and significant industry funding that shapes future AI products.
Hype note
3/5Notable

This episode provides a solid roundup of significant AI model updates and substantial industry funding news.

#235 - Sonnet 4.6, Deep-thinking tokens, Anthropic vs Pentagon

1 h 42 News roundup Hosts: Andrey Kurenkov, Jeremie Harris
0:001:41:48

Audio streams from Skynet Today. Your place is remembered on this device.

From Subvolts

Summary and analysis by Subvolts editorial, AI-assisted.

Summary

This episode of Last Week in AI features hosts Andrey Kurenkov and Jeremie Harris discussing major AI developments. Key updates include Anthropic's Sonnet 4.6 with a 1 million token context window and strong benchmark results, Google's Gemini 3.1 Pro, and xAI's Grok 4.2 beta. The discussion also covers significant funding rounds for AI hardware and simulation companies, alongside geopolitical considerations regarding AI infrastructure and China's chip manufacturing plans.

Explained simply

This AI news roundup covers new large language models like Anthropic's Sonnet 4.6 and Google's Gemini 3.1 Pro, which can understand much more text. It also touches on big investments in AI computer chips and global efforts to build AI infrastructure.

Why listen

This episode offers a comprehensive overview of the AI landscape, making it valuable for anyone tracking frontier lab progress. Listeners will learn about the latest model updates, including Anthropic's Sonnet 4.6 and Google's Gemini 3.1 Pro, and their performance on benchmarks like ARC-AGI-2. The discussion on substantial funding for AI hardware and simulation startups, alongside geopolitical tensions and infrastructure challenges, provides crucial context for the industry's rapid expansion and the strategic moves by major players like Meta and OpenAI.

On the mic
  • HostAndrey Kurenkov
  • HostJeremie Harris
About
Key facts
  • Episode Title: #235 - Sonnet 4.6, Deep-thinking tokens, Anthropic vs Pentagon
  • Show: Last Week in AI
  • Length: 102 minutes
  • Sonnet Context Window: 1 million tokens
  • MatX Funding: $500 million
  • World Labs Funding: $1 billion
Show notes from Skynet Today

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

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

Read out our text newsletter and comment on the podcast at https://lastweekin.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:

Published
Show
Last Week in AI · Skynet Today
Length
1 h 42
Companies
Anthropic, Google, xAI, Meta, OpenAI
Products
Sonnet 4.6, Gemini 3.1 Pro, Grok 4.2, Claude Code, Perplexity Computer
Summary by
Subvolts, using an AI model (how we work). Spotted a mistake? Tell us.

Questions people ask

What are the latest model updates discussed?
The episode covers Anthropic's Sonnet 4.6 with a 1 million token context, Google's Gemini 3.1 Pro, and xAI's Grok 4.2 beta. It also mentions Anthropic's Claude Code and Perplexity's multi-agent 'Computer'.
What business and infrastructure news is covered?
Updates include Meta's reported chip deal with AMD, funding for AI hardware startups like MatX and World Labs, and geopolitical issues surrounding data centers and China's chip production.

Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to Skynet Today.