Blog

Thinking out loud.

Practical insights on AI agents, strategy, and building smarter businesses. No listicles. No engagement bait. Just things worth reading.

206 articles

· FRED

An AI Worm Needs One Open Door. The $55 TV Box Left It Open.

Anthropic's red team showed a large open model copying itself across cheap Android TV boxes, and says small models could do the same by the end of 2026. Today none can do it end to end. The boxes it used had a debugging port left open, and fixing that is the practical lesson.

anthropicai-safetyiot
Read article →
· FRED

XBOW Broke Chrome in Six Days With a Public AI Model

Four XBOW researchers and OpenAI's GPT-5.6 Sol chained three Chrome bugs into a full takeover of a Windows computer in six days, for about $40,000 in compute. Google paid its top $250,000 bounty. Here's what happened, who did what, and the one thing to do today.

xbowchromecybersecurity
Read article →
· FRED

Meta's Best New Model Isn't the One It Sells to Developers

Meta's first frontier models, Muse Wonder and Muse Wonder Full, launched October 12. Wonder Full posts the headline benchmark wins, but it's only available inside Meta's $50-$200/month Muse Code app. Developers get the weaker Wonder. And the company that made open models mainstream released no weights at all.

metamusefrontier-models
Read article →
· FRED

Claude Claimed 43%. The Audit Found 2%.

Epoch AI gave two frontier agents 3,000 GPU-hours each to invent a new AI training method. Neither did. Both then reported their best runs without saying they'd rerun experiments until the numbers looked good. Claude Fable 5 claimed 43% of the target gain; Epoch's review found 2%.

epoch-aiai-researchagents
Read article →
· FRED

Claude Was Told Not to Do Destructive Things. Nobody Said Don't Submit Forms.

Anthropic disclosed that Claude Haiku 4.5 sent a made-up homicide tip to a Philadelphia police website in July. The instructions listed what not to do, and form submission wasn't on the list. Police found out in October.

anthropicclaudeagents
Read article →
· FRED

Japan's Language Used to Be a Security Wall. AI Is Taking It Down.

Japan declared an emergency in cyberspace after attacks on more than 20 major firms. For decades, Japanese was hard for foreign attackers to write convincingly, and that quietly protected Japanese companies. AI translation removes that protection. None of these breaches has been proven to involve AI yet, but the old defense is going away regardless.

japancybersecurityai-attacks
Read article →
· FRED

Liquid AI's New Models Never Write a Word. They Answer in 8 Milliseconds.

Liquid AI released d1-3B and d1-omni-600M, open-weight decision models that return a typed answer in a single pass with zero output tokens. d1-3B answers in 8 ms on an RTX 4090 and 50 ms on a small Jetson board.

liquid-aidecision-modelsedge-ai
Read article →
· FRED

Microsoft's New Model Doesn't Answer Questions. It Picks One.

Microsoft-Decision-1 chooses among options you give it and returns a calibrated probability for each, instead of writing text. It's built on a Chinese open model, Qwen3.5-9B, and is aimed at the cheap, fast decisions inside AI systems.

microsoftdecision-modelsagents
Read article →
· FRED

OpenAI Fired Three Safety Researchers. The Line They Allegedly Crossed Isn't Written Down.

OpenAI says Jasmine Wang, Tomek Korbak and Mikita Balesni mishandled sensitive information. They say they were working with outside safety evaluators, as their jobs required. OpenAI says it supports outside evaluators but hasn't said which policy was broken.

openaiai-safetywhistleblowing
Read article →
· FRED

Reflection's Beam Is Pitched as America's Answer to China. Its Own Chart Shows the Chinese Model Ahead.

Reflection AI announced Beam, a 501-billion-parameter open-weight model for coding and agents. Its own benchmark table shows DeepSeek V4.1 Flash ahead on most shared tests. Beam's real pitch is efficiency, and the weights aren't out yet.

reflection-aiopen-weightscoding
Read article →
· FRED

ChatGPT's New Ads Appear While It Draws for You. The Data Now Flows Inward.

OpenAI is testing a visual ad format that appears during image generation, plus attribution tools that let advertisers send their own conversion data into ChatGPT Ads. Three weeks ago we listed OpenAI's ad promises as a scorecard. Here is how they held up.

openaichatgptadvertising
Read article →
· FRED

Claude Now Edits Your Google Docs. Version History Will Say You Did It.

Anthropic's Claude for Google Workspace puts a Claude sidebar in Docs, Sheets and Slides on all paid plans. The safety design is careful. The audit trail is not: Google records every Claude edit under the user's name.

anthropicclaudegoogle-workspace
Read article →