Our 243rd episode with a summary and discussion of last week's big AI news!

Recorded on 04/29/2026

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:

OpenAI released GPT-5.5 with strong coding-oriented improvements, a system card discussing chain-of-thought monitorability and misalignment testing, higher pricing than GPT-5.4, and notable quirks like a system-prompt warning about “goblins.”
xAI launched Grok Voice Think Fast 1.0, claiming large benchmark leads for real-time voice agents and reporting major Starlink customer-support automation and sales conversion impact.
DeepSeek open-sourced DeepSeek V4 (Pro and Flash) featuring MoE scaling and 1M-token context via hybrid/compressed attention changes, while Tencent released Hunyuan 3 preview with weaker benchmark performance; a new long-horizon agent benchmark (Clawmark) shows low task success rates.
Major business, legal, and policy updates include Google’s planned up-to-$40B investment and 5GW compute commitment to Anthropic, Meta’s AWS Gravitron deal and China blocking Meta’s Manus acquisition, a revamped OpenAI–Microsoft agreement, ongoing Musk–OpenAI trial developments, and new safety/security research on sabotage, document degradation under delegation, and bit-flip attacks.

Timestamps:

Applications & Business
(00:53:03) Google Plans to Invest Up to $40 Billion in Anthropic
(00:56:26) Meta will use hundreds of thousands of AWS Graviton chips
(00:59:51) China blocks Meta's $2 billion takeover of AI startup Manus
(01:01:45) OpenAI shakes up partnership with Microsoft, capping revenue share payments
(01:07:13) Elon Musk Testifies of AI Risk at Trial, Says OpenAI Tried to ‘Steal’ a Charity - WSJ
(01:11:50) Judge rejects DOJ bid to delay Anthropic appeal in Pentagon dispute
(01:14:42) Google’s Gemini can now run on a single air-gapped server — and vanish when you pull the plug
(01:19:07) DeepMind's David Silver just raised $1.1B to build an AI that learns without human data | TechCrunch

Policy & Safety
(01:22:47) Evaluating whether AI models would sabotage AI safety research
(01:28:59) LLMs Corrupt Your Documents When You Delegate
(01:32:50) Temporal Sparse Autoencoders: Leveraging the Sequential Nature of Language for Interpretability
(01:39:53) Memorandum on Adversarial Distillation of American AI Models
(01:41:41) Teen boys are dating their AI chatbots—and experts warn it could kill their careers | Fortune
(01:43:57) Announcing the Anthropic Economic Index Survey
(01:45:21) Scoop: CISA lacks access to Anthropic's Mythos

Research & Advancements
(01:49:15) Maximal Brain Damage Without Data or Optimization: Disrupting Neural Networks via Sign-Bit Flips

See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

#243 - GPT 5.5, DeepSeek V4, AI safety sabotage

Last Week in AI