News archive
All posts and signals, newest first.
SearchAnthropic basically admitted Opus 5 has problems
After Chubby called out the laziness, sloppy mistakes and the downgrade from 4.6, Thariq agreed with the criticism and said improving consistency is…
Codex limits reset after hitting 20M active users
Claude Code can now be started remotely from your phone
Before, Remote Control only let you continue a session already running on your computer. Now you can choose your computer and folder, then launch a…
Codex is getting a BANKED reset
OpenAI is also investigating reports of usage limits draining way too fast.
Xiaomi MiMo V3 may be stealth-launching as Ox Alpha
Ox Alpha appeared anonymously on OpenCode and quickly hit ~80% vs Fable’s 65% in an early DeepSWE-style test. It’s almost certainly Chinese. Some…
Muse Spark 1.2 is now free on OpenCode
Easiest way to get started is OpenCode Desktop https://opencode.ai/download Power users can grab an API key here https://opencode.ai/zen - Base URL…
ChatGPT just got a quick launch for Codex Remote
Turn it on and ChatGPT opens straight into Remote, ready to work with Codex on your computer from your phone. How to set it up: ChatGPT → Remote → ⋯…
Anthropic has extended the higher Claude Code limits
This was pretty much expected, given all the backlash Anthropic has been facing lately.
Exactly 33 years ago today, a boy named Alex was born
He grew up watching computers become personal, the internet become universal, and technology reshape almost every part of our lives. And now he…
Qwen 3.8 27B beat GPT 5.6 Terra
This happened in the AA Agentic Index, a benchmark that tests how well models handle agentic tasks. And you can run Qwen 3.8 27B locally on a single…
Cursor launched its own GitHub
Origin lets you host repos directly in Cursor, open PRs, browse code, review changes, and run agents inside each repo. You can also connect GitHub…
Cursor employee is giving away a free month of $200/mo Cursor Ultra + Grok Bot
All you have to do is cancel your current AI subscription and DM her a screenshot along with your email.
A new trend is going viral on X: everyone is turning into Dario Amodei, the CEO of Anthropic
It was started by OpenAI employee Andrew Ambrosino. Here’s the thread, you can grab the prompt and drop your version there…
DeepSeek slashed OpenCode Go limits 17x
DeepSeek V4 Flash limits dropped from 63,300 to 3,800, while V4 Pro fell from 3,450 to 1,050. This comes as DeepSeek API prices have jumped sharply.
AI Watermarks Remover is out
The project hit 11K stars in just a couple of days. It works in three ways: strips hidden text characters with scripts, disrupts statistical…
Get 3 months of SuperGrok Heavy for $99
1. Grab 2 months of SuperGrok for $15 https://grok.com/supergrok/targeted-offer/email-supergrok-50-off-2mo 2. Then upgrade to Heavy for $84…
DeepSeek Harness beat OpenClaw’s record and became the fastest growing repo on GitHub
It took just 42 hours to hit 100K stars. OpenClaw hit 100K in 48 hours, but the repo was already two months old.
GPT 5.6 Luna is now a native subagent
It can now work alongside Sol and Terra. Until now, Luna was on the previous subagent system, so users had to rely on workarounds to plug it in.
Qwen hit 3 billion downloads
Alibaba’s models pulled in over 3B downloads worldwide in just six months. On Hugging Face, Qwen is running at roughly 5x Google’s download count…
Codex is getting 94% faster
This applies to long sessions. Previously, opening a huge chat meant Codex loaded almost the entire history at once. Now it will load only the part…
Alibaba released Qwen3.8’s weights
A model at Claude Opus 4.6 level, a flagship just six months ago, can now run on a single RTX 4090. Qwen3.8 27B uses around 17 GB of VRAM when…
Z.ai released GLM 5.3
The new model is noticeably stronger than GLM 5.2 in coding, agents, and cybersecurity. Across benchmarks, GLM 5.3 is already very close to GPT 5.6…
Very few people know this, but Muse Spark 1.2 might be the best value model on the market right now
Its intelligence is around GPT-5.6 Terra. But if you use Muse Code and agree to code analysis through the Contributor tier, the price drops…
Google just dropped Gemini 3.7 Flash
It’s a big upgrade over 3.6 Flash for coding and agents. Google also cut the price in half to $0.75 per 1M input tokens and $3.75 per 1M output.…
DeepSeek is not cheap anymore
V4 Pro comes with a brutal API price hike. During peak hours cached input jumps 12x, while output gets 4.5x more expensive. The dirt cheap DeepSeek…
Cursor opened its own cafe in NYC
They took over a cafe in Brooklyn for a day and invited everyone to come vibe code together. Bring your laptop, grab free coffee, code next to the…
Opus 5 could cost you 5 extra hours every month
Anthropic says the average Claude Code user spends around 20 hours a week with it. Opus 5 writes about 60% more than Fable 5. That can add up to…




























