HackrLife

GPT-6.1 Terra Halves API Prices, and the No-Signup First Task That Recruits Users

Plus: LangChain's code-enforced agent gates, the VibeVideoBench paper, Dedica reviewed, and 5 startups funded yesterday

Anthropic Files to Go Public on $16.3B in Half-Year Revenue

Plus: the read-back loop that makes voice agents get details right, a benchmark that catches agents passing the wrong way, Loqui reviewed, and 5 startups funded yesterday.

Kimi K3.1 Pro Closes In on Opus and 86% of Small Businesses Use AI

Plus: can selling to AI agents really bring you users, Shopify's hint-don't-answer loop, 7 ways RLHF goes wrong, a Notion Auto review, 3 startups funded this week, and quick hits.

ChatGPT Ads Reach 89 Countries, and "Build With AI" Hiring Demand Up 51%

Plus: how Anthropic risk-tiers its code review, an IBM paper on why RAG tricks don't stack, Radaric reviewed, 5 startups funded yesterday, and quick hits.

Flint puts a 40B Coding Model on a Laptop, and the median AI paid user pays $25 a month

Plus: why AI won't overtake search in 2026, per a 2025 EMARKETER forecast, cheap-model guardrails that rival frontier judges, Fastlane 2.0 reviewed, and 5 startups funded yesterday.

Otter's $320M Non-Equity Raise and How to Run Your Own AI Credits Programme

Plus: why more memory can make agents worse, Meta's thin-orchestrator pattern, Onepin for AI voiceovers, and 5 startups funded yesterday.

Only 13% of CIOs say AI has paid off

Automated evals that lift accuracy from 78.6% to 90.5% and an open model for text-heavy graphics

Meta Rewrites Its Crawler Terms, the 30-Day Window Where Upsells Work Up to 10x Better, and a Router That Cut Median Agent Cost 64%

Plus: Later's 89% jump in creator payouts, why coding agents fix more bugs with whole files, ngram 2.0 reviewed, and 6 startups funded since Friday.