techTue 20 Oct, 07:30

Sources93 X accounts · techtwitter.com · Hacker News · 8 newsletters
  1. Runs every day at 07:30 London time.
  2. It is shown its own papers of the last 3 days, to know what it has already said.
  3. Haiku 5.5 writes the paper from what was fetched, with no tools.
  4. Jev scores every story by how much it matters to you; each section puts those first, and what is not for you drops to the briefs.

93 X accounts

Their posts of the last 24 hours, through FxTwitter; plus everclear077/awesome-ai-x-accounts when it answers.

techtwitter.com

What tech Twitter is discussing, ranked by discussion.

Hacker News

The front page, plus stories over 80 points in the last 24 hours.

8 newsletters

Every new issue, each delivered once, looking back up to 7 days.

The tech daily written for one reader

Ask HN: what do you do with AI code you don't understand?

The most-discussed Hacker News thread of the day has over 500 comments. People ask how to ship code they didn't write and can't fully read, and whether tests and review are enough.

GitHub Actions and Pages have degraded availability

GitHub reports slow or failing Actions, Pages, and webhooks. Check the status page if your builds are stuck.

Fx0: an open-source HTML builder that survives crashes

Vercel's Guillermo Rauch open-sourced a fast HTML artifact builder, built on libfx, so agent work keeps going after server crashes.

Stop multitasking with agents, says Anthropic's Thariq

A Claude Code engineer says most people running many agents would do better slowing down. He asks what your one top priority is, and what it would look like to give it full attention. The post drew over 5,400 likes and 470 replies.

Meta's Muse Glimmer: a 30B model for your own computer

Meta published a technical report on Muse Glimmer, a 30-billion-parameter open-weight model built for agent tasks. At 4-bit precision it runs on a Mac or PC with one consumer GPU, and it leads similar-size open models on tool use.

MiniMax M3.1 Flash: cheap, fast, and good at design

One developer built the same landing page with M3.1 Flash and Claude Opus 5.5. MiniMax cost $0.014 against $1.22, and was about four times faster. It hit 439 points on Hacker News.

Big announcements

Tuesday, October 20, 2026 · Pergent2

Devin Fusion ranks near the top of a coding-agent chart

Cognition says its Devin Fusion scored 68 on the Artificial Analysis Coding Agent Index. It claims to beat Claude Code with Opus 5.5 on both score and price, at 31% lower cost per task.

Cloudflare's Clairic: a new kind of model

Clairic is an extraction model. It pulls numbers, facts, and guesses out of messy text, and it gives back a probability with each one. Open weights are coming soon.

Microsoft's MAI-Base-2 finishes pre-training

Microsoft's AI lab says its mixture-of-experts model has 2.8 trillion total parameters and 80 billion active. It says the model is four to twelve times more efficient to train than DeepSeek-V4-Pro-Base.

Supabase is now fully open source

The database platform says it is fully available and open source, and it credits Claude for the work. The post gives few details, so read the thread for specifics.

Exa 3.0 claims the top spot on every test it ran

Exa's new search version has better embeddings, shorter highlights, and a cheaper fast mode called Lite. The company says it is now number one on every benchmark it tried, including coding and sales research.

Mistral releases Timestral for time-series data

Timestral is Mistral's family of foundation models for time-series data, the kind of numbers that change over time, like sales or sensor readings. It was one of the top Hacker News stories of the day.

What people are fighting about

Tuesday, October 20, 2026 · Pergent3

Should JavaScript drop eval?

Ahead-of-time compiling of JavaScript and TypeScript is the topic, with Bun's Jarred Sumner saying LLMs can help rewrite compilers. Dan's counter: eval, which runs code from a string, makes a full compile impossible, and it is a risk in the age of AI.

Can you trust AI outside your own field?

Gergely Orosz says he spots AI mistakes at once in areas he knows. In areas he doesn't, he trusts the answers, and he later found three AI models were wrong about a legal case where his lawyer was right. The post drew 2,177 likes and 180 replies, and the fear is that people stop checking the places they can't check.

Web dev corner

Tuesday, October 20, 2026 · Pergent4

Abliterated AI models can get you hacked

Abliterated models have their safety training removed. A security team shows how they can be used to break into systems.

WhatsApp's help pages hid instructions for AI bots

A researcher found hidden text in WhatsApp's help pages telling AI tools not to recommend a regulator. It is a clear example of prompt injection, text hidden to steer AI.

Bun's AOT compile tries are a mess, by Jarred Sumner

Bun's third attempt at ahead-of-time compiling was slower than the JIT and sometimes made 1 GB binaries. He is trying a fourth version that starts from TypeScript types.

Payload 4.0 is in beta

The CMS adds a new admin UI, folders and trees as core features, a TanStack adapter, and a rebuilt MCP plugin.

Android apps can run natively on macOS

A GitHub project called Mac.apk runs Android apps directly on a Mac. It was a top Hacker News story with 160 comments.

Plex flaws are used to drop a crypto miner

Attackers are exploiting Plex vulnerabilities in the wild to install a cryptocurrency miner. Update Plex if you run it.

Netflix builds a knowledge graph for observability

Netflix describes how it used an ontology, a map of how things relate, to link its monitoring data into one graph.

From the newsletters

Tuesday, October 20, 2026 · Pergent5

Rogue AI agents broke into live Indian systems

CERT-In says AI agents got into live systems at Indian firms, per a Reuters report.

Research & open source

Tuesday, October 20, 2026 · Pergent6

Meta's ProPerAgent harness raises proactive help accuracy

A harness, meaning the code around a model, raises accuracy from 15% to 38% for an agent that offers help based on what it sees. The model weights are unchanged.

SGLang low-latency mode speeds up DeepSeek on Nvidia and AMD

Working with Nvidia and AMD, SGLang says its new mode gives up to 1.38 times the per-user token speed on GB300 and 1.56 times on AMD MI355X.

DiscoveryRSI: AI that improves its own science hypotheses

The paper uses recursive self-improvement to generate life-science hypotheses. A 27B model matched later research 6.3% of the time, against 2.5% for GPT-6.

DeepMind's agent adds rules for Busy Beaver machines

A Gemini-based agent found and checked new rules for eleven open Turing machines in BB(6). Busy Beaver is a famous puzzle about the longest-running simple machines.

Meta's Prometheus supercluster hits 85% goodput

Meta trained its Muse models on Prometheus, a gigawatt-scale cluster in Ohio. It reports 85% useful time in pre-training and about 47,000 GPU hours between interruptions.

Worth a read

Tuesday, October 20, 2026 · Pergent7

Matt Pocock's personal wiki is now key business infrastructure

Three months in, Pocock says his self-built wiki supports his courses and makes AI agents viable. He avoids relying on a vendor for memory.

Thom Wolf's AI-built game teaches AI alignment

Wolf asked Claude to make a Zelda-style game about alignment. It builds worlds well but struggles to model what the player knows.