Anuar
@_startuphacker
CTO @ http://tyce.ai - AI agent for documents | my journey from Kazakhstan's steppes to building AI B2B SaaS
قد يعجبك
We just launched! And we are #1 on ProductHunt! 🚀 Tyce is an AI Agent that crafts the smartest documents. With Tyce you can write in the flow, Cursor-like vibes but for knowledge workers. Tyce is deeply integrated with our Word Processor, it can find relevant information in…
Ilya Sutskever is a Carl Sagan of AI
The @ilyasut episode 0:00:00 – Explaining model jaggedness 0:09:39 - Emotions and value functions 0:18:49 – What are we scaling? 0:25:13 – Why humans generalize better than models 0:35:45 – Straight-shotting superintelligence 0:46:47 – SSI’s model will learn from deployment…
Hack: if the LLM is not following your system prompt instruction, just append that instruction to the latest user message and it would work! LLM didn't follow your system instruction because there was some contradictory information in your message history. Appending to the…
Considering to live in our customer’s office, they have a nice view
You can sustain infinite amount of pain if you convince yourself that future is bright
In final form our idea is so scary to our users that they will oppose the buying process of our product 🤡
We landed a customer who built this building 👇 Feels good working with a customer who appreciates beauty
Prediction: Next year Apple will release human-verification features whenever you use iPhone / Macbook to disambiguate if it's you vs AI using your devices. This will be used in social apps like Instagram to show human authentic content.
Best Models from API perspective: Reasoning: GPT-5 reasoning high Long context: GPT-4.1, Gemini Tool Calling: Claude Sonnet 4.5 Instruction following: Claude Sonnet 4.5 Image input: Gemini 2.5 Pro PDF: Gemini 2.5 Pro In-built tools (Web Search, File Search, Computer Use): OpenAI
Updated my AI model capabilities table. Here's my favorite model for each task: 1. Everyday answers: I use ChatGPT because GPT-5 is fast, concise, and almost never hits rate limits. 2. Writing: Claude is my main editor. I give it examples of my best writing and it just gets my…
I like @TraycerAI idea of converting a github issue to an implemented feature or bug with AI. The problem is that there is not enough information in the issue description to fully give context to the LLM. For QA bugs, if it was possible to fully reproduce the user flow - take…
One of 2025 Nobel Prize in Physics laureates is John Hopfield, who showed how AI can be formed from the laws of physics! He discovered the Hopfield Network. It’s a form of a Recurrent Neural Network (RNN) in AI that can remember and recall a memory, inspired by energy…
listen to customer calls, pause grinding operations to strategize and push for more
AGI is just around the corner but somehow my life hasn’t changed much 🤡
All it takes is 1 line change in your system prompt and your users run for their lives. Hence everyone complaining intermittently about Cursor, Codex and Claude Code… and switch between them. With every release there are new “prompt bugs” that evade existing evals and land in…
Should be shown before any election and taught in schools.
Introducing Vancouver Spends See how the city's revenue and spending affect Vancouver residents and programs. Thanks to Boris Tsao for leading this end-to-end: canadaspends.com/en/vancouver
United States الاتجاهات
- 1. Caleb 55.9K posts
- 2. Bears 83.1K posts
- 3. Packers 69.9K posts
- 4. Jeff Kent 6,317 posts
- 5. Texans 21K posts
- 6. #BaddiesUSA 6,368 posts
- 7. #GoPackGo 12K posts
- 8. Notre Dame 171K posts
- 9. Ben Johnson 7,018 posts
- 10. #ChiefsKingdom 3,404 posts
- 11. DJ Moore 3,281 posts
- 12. Collinsworth 1,669 posts
- 13. Raiders 36.1K posts
- 14. Nixon 14.6K posts
- 15. Mattingly 3,685 posts
- 16. Mattingly 3,685 posts
- 17. Nico Collins 1,532 posts
- 18. Kmet 1,789 posts
- 19. Shedeur 117K posts
- 20. ESPN 124K posts
Something went wrong.
Something went wrong.