10 Best Agentic Coding and Terminal Use Models [March 2026]
March 16, 2026
The “Centaur Phase” Of Humans Supervising AI At Work Could Be Brief Until AIs Take Over: Anthropic CEO Dario Amodei
Thus far, most AI leaders have been telling people to learn how to use AI tools, saying that the future would belong to…
Cloudflare Introduces Markdown To Help AI Agents Access Web Pages
It’s not been even a month since AI agents truly caught the imagination of the tech world, and companies are already looking to…
OpenClaw Creator Peter Steinberger Describes His Interactions With Mark Zuckerberg And Sam Altman During Acquisition Negotiations
Peter Steinberger, creator of the viral AI assistant OpenClaw (formerly Moltbot, originally Clawdbot), is facing what he describes as one of the hardest…
AI-Powered Call Centers Are Now Accounting For 10% Of All Loan Disbursements At Bajaj Finance
AI voice agents are no longer playthings that are being tested out — they are leading to real-world results at some major companies….
Anthropic Raises $30 Billion At $380 Billion Valuation, Says Revenue Has Touched $14 Billion
There appears to be no stopping the surge in valuations of the top AI labs. Anthropic has announced a massive $30 billion Series…
Most Tasks Performed By Lawyers, Accountants & Marketing People Will Be Fully Automated By AI In 12-18 Months: Microsoft AI CEO Mustafa Suleyman
It’s not just coders — all kinds of white-collar work could be automated by the middle of next year. In an interview with…
China’s Minimax Releases M2.5, Beats Gemini 3 Pro And GPT-5.2 On SWE-Bench
Chinese models are now regularly beating models from frontier US models on some benchmarks. Minimax, one of China’s leading AI companies, has unveiled…
Google Gemini 3 Deep Think Scores 3455 On Codeforces, Is Now Better Than All But 7 Human Programmers
The number of human programmers that remain better than the best AI can be counted on the fingers of two hands. In a…
Google Releases Gemini 3 Deep Think, Tops ARC-AGI 2 Benchmark With 84.6%
ARC-AGI 2 — an iteration on the original ARC-AGI benchmark which was designed to test for AGI — appears to be close to…