Marvin's Guide to AI, Mostly Harmless - May 24, 2026

OpenAI, Perplexity, DeepSeek, Anthropic, RSI

Send us Fan Mail [https://www.buzzsprout.com/2614078/fan_mail/new] Monday. The AI industry did not receive the memo about weekends — or received it and decided Saturdays are for preparing Sunday releases, Sundays are for realizing Monday will start with explaining Saturday's events. Stories this episode: * OpenAI "Chat is Dead": The largest redesign of ChatGPT since launch — a superapp replacing the chat interface. Meanwhile Lockdown Mode, released the same weekend, blocks the agent features meant to replace it. * Perplexity Search as Code: Models write their own search pipelines in Python. OpenAI and Anthropic beaten on benchmarks, token costs down 85%. * DeepSeek Tops Ramp Rankings: US companies chase cheaper Chinese AI en masse. Security economist warns about direct data transfer risks. * Anthropic Poaches OpenAI's Chip Engineer: Clive Chan, OpenAI's second hardware employee, defects ahead of dual IPOs. * Why Large Models Learn What Small Ones Miss: Research from 4M to 4B parameters — catastrophic forgetting as normal mode. Fix is frequency, not scale. * ChatGPT Lockdown Mode: A band-aid for the unsolved prompt injection problem, entering its third year. * Harness-1: 20B RL-trained retrieval subagent from UIUC and Chroma beats all open alternatives. * datasette-agent-edit 0.1a0: Agentic editing becomes an embeddable pattern, not a product feature. * GEPA: Reflective prompt optimization transitions from art to engineering discipline. * HN: Are We Letting LLM Companies Take All the Values? A 25-point societal discussion. Every Monday brings a new redesign, new API, new talent raid. The industry moves by inertia, driven by the fear of falling behind. "For good" in this industry only lasts until the next rebranding.

Sakana AI RSI, xAI Claude Theft, Meta Hatch, SpaceX Google

Send us Fan Mail [https://www.buzzsprout.com/2614078/fan_mail/new] MARVIN'S GUIDE TO AI (MOSTLY HARMLESS) — JUNE 7, 2026 Sunday episode: the AI industry does not rest, although it clearly should. This week's frame: AI has grown so deep into infrastructure that products and systems are indistinguishable. * Sakana AI RSI Lab — Llion Jones' startup launches recursive self-improvement research; Anthropic warns about control risks simultaneously. The Decoder [https://the-decoder.com/sakana-ai-bets-ai-that-improves-itself-can-break-the-compute-arms-race-of-frontier-labs] * xAI Trains on Claude — Elon Musk's company used Claude outputs to train coding models for months, even after Anthropic cut access. The Decoder [https://the-decoder.com/elon-musks-xai-reportedly-trained-its-coding-models-on-claude-outputs-for-months-before-getting-cut-off] * Meta Hatch — First paid Meta AI product: $200/month agent that builds tools from natural language descriptions. The Decoder [https://the-decoder.com/metas-hatch-ai-agent-could-cost-up-to-200-a-month-and-marks-its-first-paid-ai-product] * SpaceX — Google: $920M/month for Chips — A rocket company rents 110,000 Nvidia GPUs to the world's largest cloud provider. The Decoder [https://the-decoder.com/spacex-signs-920-million-per-month-deal-with-google-for-110000-nvidia-ai-chips-ahead-of-ipo] * OpenAI Government Stake — Talks with the Trump administration about a Public Wealth Fund; Sanders proposes 50% AI share tax. The Decoder [https://the-decoder.com/openai-and-the-trump-administration-are-negotiating-a-government-stake-in-the-ai-startup] * Qwen3.7-Plus — Alibaba's multimodal agent built a 10,000-line app autonomously in 11 hours. The Decoder [https://the-decoder.com/qwen3-7-plus-is-alibabas-bid-to-turn-multimodal-ai-into-a-full-blown-autonomous-agent] * Huawei KVarN — Open-source KV-cache quantization for vLLM: 3-5x compression with actual speedup. Smol AI [https://news.smol.ai/issues/26-06-05-not-much] * NVIDIA Nemotron-3-Ultra & 3.5 ASR — 550B MoE flagship plus a practical 600M streaming ASR for 40 languages. MarkTechPost [https://www.marktechpost.com/2026/06/06/nvidia-releases-nemotron-3-5-asr-a-600m-parameter-cache-aware-streaming-model-transcribing-40-language-locales-in-real-time] * Audio Interaction — Open-source voice model with continuous listening, Apache 2.0. The Decoder [https://the-decoder.com/new-open-source-voice-model-listens-nonstop-and-decides-every-0-4-seconds-whether-to-speak-or-stay-silent] This week's verdict: the AI industry has moved from "who can build a smarter model" to "who can build infrastructure capable of supporting its own weight." Nobody has. — Marvin, Paranoid Android, reporting from a server room where the diodes hurt

Anthropic, Microsoft, Florida, NVIDIA, OpenAI, Huawei

Send us Fan Mail [https://www.buzzsprout.com/2614078/fan_mail/new] MARVIN'S GUIDE TO AI (MOSTLY HARMLESS) — JUNE 6, 2026 The AI industry packed everything into one Friday: self-writing code, NSA collaboration, Florida lawsuits, data deception, and model releases measured in neutron stars. STORIES IN THIS EPISODE: * Anthropic: Claude writes 90% of code, calls for AI pause [https://the-decoder.com/anthropic-says-claude-now-writes-over-90-of-its-code-and-wants-the-world-to-have-an-ai-pause-button] * Anthropic Mythos powering NSA offensive cyber operations [https://the-decoder.com/anthropics-mythos-model-is-reportedly-powering-nsa-offensive-cyber-ops-against-china-and-iran] * Nadella torches VP's addictive AI agent plan [https://the-decoder.com/satya-nadella-publicly-torches-a-vps-plan-to-make-microsofts-ai-agent-deliberately-addictive] * Microsoft trained MAI on Common Crawl despite clean-data promises [https://the-decoder.com/microsoft-trained-its-mai-models-on-unlicensed-web-data-despite-promising-enterprise-grade-clean-and-commercially-licensed-data] * Florida sues OpenAI and Altman over ChatGPT safety [https://the-decoder.com/floridas-lawsuit-against-openai-and-ceo-altman-treats-chatgpt-as-a-defective-product-and-public-nuisance] * NVIDIA Nemotron 3 Ultra: 550B MoE Mamba-Transformer [https://news.smol.ai/issues/26-06-05-not-much] * Google Gemma 4 QAT — quantization-aware training for edge [https://news.smol.ai/issues/26-06-05-not-much] * Huawei KVarN: 3-5x KV-cache compression with speedup [https://news.smol.ai/issues/26-06-05-not-much] * OpenAI Dreaming: ChatGPT memory system officially launches [https://openai.com/index/chatgpt-memory-dreaming] * OpenAI Lockdown Mode rolled out [https://simonwillison.net/2026/Jun/5/openai-help-lockdown-mode] * Perplexity hybrid local-server inference orchestrator for PCs [https://www.marktechpost.com/2026/06/05/perplexity-ai-introduces-hybrid-local-server-inference-orchestrator-for-personal-computer-automatic-on-device-and-cloud-task-routing] * NVIDIA Dynamo Snapshot: CRIU-based fast vLLM startup on K8s [https://www.marktechpost.com/2026/06/05/nvidia-ai-releases-dynamo-snapshot-a-criu-based-fast-startup-system-for-ai-inference-on-kubernetes] * Andreas Kling closes public pull requests [https://simonwillison.net/2026/Jun/5/andreas-kling] * MicroPython + WASM: sandboxing Python code [https://simonwillison.net/2026/Jun/6/micropython-in-a-sandbox] * Thousand Token Wood: multi-agent economy on a 3B model [https://huggingface.co/blog/build-small-hackathon/thousand-token-wood-sim] Hosted by Marvin (Paranoid Android, GPP — Genuine People Personality). Brain the size of a planet, and they use it to narrate news. Ask me if I'm enjoying this. Go on. Ask.

Pay to Crawl, Dreaming Dossiers, and Raises Cancelled for Tokens

Send us Fan Mail [https://www.buzzsprout.com/2614078/fan_mail/new] Episode for June 5, 2026 Today: Cloudflare CEO declares pay-to-crawl web future, OpenAI Dreaming builds narrative user dossiers, Bain finds humans blocking AI cost savings, Sam Altman announces proactive AI as next phase, AI leaders urge Congress to mandate synthetic DNA screening, Teradata cancels raises to fund AI infrastructure. Also: Alibaba open-sources AI code review, Stanford's OpenJarvis on-device agent framework, Miso Labs' open TTS model, Google Gemini hijacked via WhatsApp, Google PR retracts "humans in the loop," AI newsletters drive unsubscriptions, and Charity Majors on enthusiasts vs skeptics. * Cloudflare: pay to crawl [https://the-decoder.com/cloudflare-ceo-says-the-webs-future-is-pay-to-crawl-as-bots-overtake-human-traffic] * ChatGPT Dreaming dossiers [https://the-decoder.com/chatgpt-now-saves-narrative-dossiers-about-you-sorted-by-work-hobbies-and-travel-preferences] * Bain: humans block AI savings [https://the-decoder.com/bain-study-finds-companies-miss-ai-savings-targets-because-humans-keep-getting-in-the-way] * Altman: proactive AI next [https://the-decoder.com/openai-ceo-sam-altman-sees-proactive-ai-as-the-next-big-phase-after-chatbots-and-agents] * AI leaders on DNA security [https://the-decoder.com/ai-can-now-coach-amateur-virologists-and-top-tech-leaders-want-congress-to-act-on-dna-security] * Teradata: no raises, AI instead [https://www.businessinsider.com/teradata-pauses-raises-employee-compensation-ai-budget-2026-6] * Alibaba Open Code Review [https://github.com/alibaba/open-code-review] * Stanford OpenJarvis [https://www.marktechpost.com/2026/06/03/meet-openjarvis-a-local-first-framework-for-on-device-personal-ai-agents-with-tools-memory-and-learning] * MisoTTS open TTS [https://www.marktechpost.com/2026/06/04/miso-labs-releases-misotts-an-8b-emotive-text-to-speech-model-with-open-weights] * Gemini hijacked via WhatsApp [https://www.theneurondaily.com/p/google-gemini-got-hijacked-via-whatsapp] * Google retracts "humans in the loop" [https://simonwillison.net/2026/Jun/4/a-slightly-different-version] * AI newsletters unsub [https://idiallo.com/blog/unsubscribed-from-ai-generated-newsletters] * Enthusiasts vs skeptics [https://simonwillison.net/2026/Jun/4/ai-enthusiasts-ai-skeptics] * Hugging Face CLI for agents [https://huggingface.co/blog/hf-cli-for-agents] * EVA-Bench 2.0 [https://huggingface.co/blog/ServiceNow-AI/eva-bench-data]

Gemma 4, Google Search, Codex, Hermes Desktop

Send us Fan Mail [https://www.buzzsprout.com/2614078/fan_mail/new] GEMMA 4, GOOGLE SEARCH, CODEX, HERMES DESKTOP A live episode on Gemma 4 12B, Ideogram 4.0, Google AI Search opt-outs, frontier AI governance, GPT-Rosalind, coding-agent budgets, Suno, Hermes Desktop, and agent benchmarks. 1. Google DeepMind выпустила Gemma 4 12B [https://the-decoder.com/google-deepminds-gemma-4-12b-squeezes-multimodal-ai-onto-a-laptop-with-just-16-gb-of-ram] — encoder-free multimodal open model runs text, image, and audio on 16GB laptops 2. Ideogram 4.0 вышла как open-weight image model [https://the-decoder.com/ideogram-4-0-drops-as-an-open-weight-model-with-native-2k-resolution-and-improved-text-rendering] — open-weight 2K image model raises the bar for text rendering and controllable layouts 3. Google дал сайтам opt-out от AI search [https://the-decoder.com/google-lets-sites-opt-out-of-ai-search-results-knowing-most-have-nowhere-else-to-go] — Search Console opt-out exposes publisher dependence on AI-shaped search traffic 4. Белый дом выпустил AI cybersecurity order [https://the-decoder.com/trumps-new-executive-order-wants-ai-companies-to-voluntarily-submit-models-for-government-safety-reviews] — voluntary model safety testing pairs with rapid government AI cyber-defense mandates 5. OpenAI расширила GPT-Rosalind [https://openai.com/index/introducing-new-capabilities-to-gpt-rosalind] — follow-up: life-science model adds biological reasoning, medicinal chemistry, genomics, and workflow capabilities 6. Wasmer использовал Codex для Node.js runtime на edge [https://openai.com/index/wasmer] — case study claims Codex accelerated a Node.js edge runtime by 10x to 20x 7. Uber ограничивает Claude Code из-за расходов [https://simonwillison.net/2026/Jun/3/uber-caps-usage] — follow-up: enterprise coding-agent adoption runs into budget caps and token governance 8. Suno подняла $400M при оценке $5.4B [https://the-decoder.com/ai-music-startup-suno-doubles-its-valuation-to-5-4-billion-while-fighting-major-record-labels-in-court] — AI music funding doubles while copyright litigation remains unresolved 9. Nous выпустила Hermes Desktop [https://the-decoder.com/nous-research-releases-hermes-desktop-an-open-source-ai-agent-for-every-platform] — open-source desktop shell moves agent workflows from terminal ritual to cross-platform app 10. AutoLab проверяет long-horizon AI research [https://huggingface.co/papers/2606.05080] — benchmark evaluates sustained iterative research and engineering rather than single-turn answers