When you said that you don't do policy or safety talks, I knew I liked you – this is a way for people who are not interested in AI to talk about AI and sound smart
@petergostev
-
Terminus 2 Agent Performance Below 5pta Benchmark
By
–
It is also 5pta below Terminus 2, which is meant to be the simplest agent without any tricks
-
Anthropic’s Document Assembly Capabilities Debate
By
–
Without knowing the details, my best guess is that it is just not that good at ppt/excel – a lot of these tasks are not blocked by model smarts but just how well it can put a document together. Anthropic invested in this ahead of others. Useful capability but not load bearing
-
Claude Max subscription limitations outside code environment
By
–
This is why you can't use your max sub outside of claude code
-
We Have the Technology to Solve This
By
–
We have the technology to solve this https://t.co/nDI5xLV4EB pic.twitter.com/VyoGYTS1dF
— Peter Gostev (@petergostev) 19 février 2026We have the technology to solve this
-
Nvidia Added to Log and Linear Analysis
By
–
I've added Nvidia to this (log & linear) https://t.co/3tzssfPbVz pic.twitter.com/aH55ezDGNZ
— Peter Gostev (@petergostev) 15 février 2026I've added Nvidia to this (log & linear)
-
Research Sub-Agent Based on GPT-5.2-Pro for Workflow
By
–
Do people do this? My workflow sometimes is to ask Gpt-5.2-pro in chatgpt to research something and then I share it with Codex, which is quite clunky. I appreciate the Codex search feature but it is not close to 5.2-Pro. Maybe some kind of research sub agent based on 5.2-pro
-
GPT-4 vs GPT-4o: Costly API Pricing Mistake Discovery
By
–
I remember the days when you discover that you've accidentally set your API call to be 'GPT-4' instead of 'GPT-4o' and your heart drops (because GPT-4 was like 10x more expensive than GPT-4o)
-
Using o1-pro to Show Premium AI Capabilities
By
–
If I want to impress someone, I don't use Opus 4.5 or GPT-5.2-xhigh, I use o1-pro to show that no expense has been spared