Gemini adopts Jev's Noul, Anthropic ports to AMD, DeepMind proves Signal
Every URL the pipeline pulled into ranking for this issue — primary sources plus the supporting and contradicting findings each Researcher returned. Inline citations in the issue point back here.
Sources
Gemini Decisions API simonwillison.net
Gemini Decisions API Google join Jev and OpenAI in the decision model category with “Gemini Decision 1”. This new model API is almost identical to Jev: it supports the same three types of questions (yes/no, choice, and rubric-scoring) and even reuses Jev’s term “Noul” for a yes/no question. It’s priced the same as the GPT-6 Luna decision model: 10 cents per million input tokens, and output is free. Jev remains the cheapest of the three, at 4.2 cents per million. Jev is text (and JSON) only. Ope…
Google DeepMind just mathematically verified key parts of Signal arstechnica.com
The four-week project hints that AI could make formal verification far more practical.
Anthropic’s IPO roadshow video simonwillison.net
Anthropic’s IPO roadshow video Anthropic kicked off their IPO roadshow yesterday, and their roadshow video has some previously unreleased information about how their business works. The section on “Approach to compute” with co-founder Tom Brown is particularly interesting, describing how Anthropic train and serve Claude across AWS Trainium, Google TPUs, NVIDIA GPUs, and AMD GPUs. I’m not sure why there’s no official video transcript, but here’s one that I extracted using MacWhisper. Highlights…
Token Pooling-Aware Training for Multi-Vector Embedding Models huggingface.co
Hugging Face details a training recipe that bakes token pooling directly into multi-vector embedding models, closing the quality gap that normally appears when pooled vectors replace full token sets at inference. The approach keeps retrieval accuracy while shrinking index size for ColBERT-style systems.
aws-emulator simonwillison.net
AWS released an official local emulator for its services, removing the need for paid API keys during testing. The Rust codebase landed 29,684 commits in three weeks via an agent swarm, with Swami Sivasubramanian detailing the build on Werner Vogels’ blog. Docker is required on macOS and Windows.
llm 0.37.2 simonwillison.net
Simon Willison’s LLM tool shipped 0.37.2 to fix two advisories from researcher Daniel de Araujo. The extra-openai-models.yaml mechanism for third-party OpenAI-compatible endpoints could forward the user’s real OpenAI key when a named key was unset or when responses-mode models ran with chat_completions enabled.
JioStar’s chief architect on AI discovery at 500M-user scale openai.com
JioStar’s Vijay Seshadri walks through how the Indian streamer runs AI-powered content discovery for 500 million users, covering evaluation pipelines, guardrails, and the operational lessons from moving OpenAI models into production at that scale.
Browse tools.simonwillison.net by tag simonwillison.net
tools.simonwillison.net now sports tag-cloud navigation across its 244 tools, grouped by JavaScript library, web API, and CDN host. Willison built it with Fable 5.2 credits and tree-sitter code analysis, surfacing stats like 92 tools using ES modules and 69 using the Clipboard API.
References
TechRadar — on Stella Laurenzo (AMD AI Director) GitHub critique techradar.com
Claude cannot be trusted to perform complex engineering tasks — based on analysis of nearly 7,000 coding sessions, with ‘thinking depth’ dropping 67–75% between late January and March.
AI Central Resources — technical scope caveat on the AMD port aicentralresources.com
The ‘three-day’ timeline reportedly applied specifically to ‘prompt processing’ for a single model, rather than a full, production-ready training and inference stack.
MLQ.ai — METR audit of Anthropic R&D acceleration mlq.ai
METR found a 1.9x acceleration in capability growth since October 2025, with ~80% credited to AI-assisted R&D, though the 95% confidence interval spans 1.3x to 3x and Anthropic redacted much of the most informative evidence.
Veracode 2026 GenAI Code Security Report veracode.com
44% of AI-generated coding tasks resulted in code with known vulnerabilities (CWEs) — ‘syntax is solved, but security is not.’
The SaaS CFO — Anthropic revenue-per-compute-dollar thesaascfo.com
Anthropic’s computing efficiency ratio actually dipped from 2.08x to 1.72x in Q3 2026, suggesting gains may not be as steady as the roadshow video implies.
TheAIWorld.org — on Anthropic’s S-1 RSI disclosure theaiworld.org
The amended S-1 warns Anthropic may ‘imminently’ meet the threshold for ‘dramatically accelerated progress attributable to AI’… critics dismiss the warning as a calculated ‘IPO stunt’ designed to manufacture a narrative of inevitability.
Matthew Green — Cryptography Engineering blog blog.cryptographyengineering.com
AI that can flawlessly verify code could make software ‘too secure’ — intelligence agencies that rely on exploit-based surveillance may go dark, which will restart aggressive backdoor and exceptional-access debates.
Cortist (HN-style technical recap) cortist.com
AlphaProof Code reintroduced two historical libsignal flaws — a weak-state handshake bug that let a mangled handshake create a second session with the same secret, and a missing small-subgroup check in the PQXDH responder — and produced machine-checked proofs that the modified binaries were insecure.
AI Weekly technical digest aiweekly.co
The 6.5M lines of proof rest on ~26,000 lines of human-written ‘trusted specifications’; the verification covers only one-to-one PQXDH + Triple Ratchet sessions and excludes group messaging, key generation, side channels, and the mobile app.
MLQ.ai — formalization findings mlq.ai
Formalizing Signal’s security goals into Lean surfaced 15 gaps or bugs in previously published computational analyses of the Double Ratchet — flaws in the theoretical models that expert human cryptographers had missed.
Cohn-Gordon et al. / Tamarin post-quantum Signal analysis (ResearchGate) researchgate.net
Prior Tamarin and ProVerif work verified Signal’s X3DH and Double Ratchet at the abstract protocol level under the Dolev-Yao ‘perfect cryptography’ assumption — never at the compiled machine-code level.
VentureBeat venturebeat.com
Google reported a calibration error of 0.0250, nearly half that of its closest rival, Jev
SuperpowerDaily superpowerdaily.com
The Artificial Analysis Decision Index v2.1 ranked the model 9th out of 18 available decision models at launch… Gemini achieved perfect accuracy on only 33% of documents, trailing competitors like Drex Max (63%)
Vercel AI SDK docs (ai-sdk.dev) ai-sdk.dev
experimental_decide function… allows developers to swap between specialized models like TypeSafe AI’s Jev, OpenAI’s Decisions API, and Google’s Gemini Decision 1 with minimal code changes
TypeSafe Jev docs — Noul primitive docs.typesafe.ai
‘Noul’ is a portmanteau for Bernoulli; the probability itself serves as the measure of certainty, where a value of 0.5 indicates maximum ambiguity
remio.ai — ‘OpenAI Decisions API copies Jev’s fast decisions but agent control is the real test’ remio.ai
copycat models from giants like Google and OpenAI often miss the ‘hard part’: Reinforcement Learning for Calibrated Decisions (RLCD)