Michael Lynch (@michael_irl) 's Twitter Profile
Michael Lynch

@michael_irl

ID: 1932832344437563392

calendar_today11-06-2025 16:09:54

0 Tweet

0 Takipçi

7 Takip Edilen

Garry Tan (@garrytan) 's Twitter Profile Photo

Hire the engineer who worked on the best product Hire the marketer and sales person who managed to get the worst product to outsell the best product

Deedy (@deedydas) 's Twitter Profile Photo

Software engineers who don’t find LLMs for coding useful either 1. Used it a >2mos ago and formed an opinion, before Claude Code 2. Program in an esoteric language or framework, including low level systems / C 3. Work on large preexisting codebases

Software engineers who don’t find LLMs for coding useful either

1. Used it a >2mos ago and formed an opinion, before Claude Code
2. Program in an esoteric language or framework, including low level systems / C
3. Work on large preexisting codebases
Michael Lynch (@michael_irl) 's Twitter Profile Photo

China has done the world a big favor by releasing these models freely and openly. Releasing model weights allows for the wealth created by advancements in AI to be distributed more evenly across society. Thank you China! 🐲🇨🇳

Michael Lynch (@michael_irl) 's Twitter Profile Photo

So, is there anything Cursor could have done differently to avoid losing market share to Claude Code? How would you compare the two tools? Pros and cons? I have noticed a lot of commentary about Anthropic subsidising CC as a loss leader. How do we think about that playing out

So, is there anything Cursor could have done differently to avoid losing market share to Claude Code?

How would you compare the two tools? Pros and cons?

I have noticed a lot of commentary about Anthropic subsidising CC as a loss leader. How do we think about that playing out
Michael Lynch (@michael_irl) 's Twitter Profile Photo

🚨🚨🚨 This is the best article on building AI agents in 2025. 🚨🚨🚨 Read my thread to get the summary :) anthropic.com/engineering/bu…

Michael Lynch (@michael_irl) 's Twitter Profile Photo

Ok so Cursor can now re-create entire websites (almost?) perfectly using MCP. How do we think this affects the future of frontend focused AI tools like Loveable and Bolt??? Cursor Anton Osika – eu/acc bolt.new Credit to Fili for the video

Michael Lynch (@michael_irl) 's Twitter Profile Photo

Does anyone else feel like the constant “warnings” from tech CEOs about AI are actually just marketing statements for their product? And from the journalist’s point of view, the article practically writes itself …

Michael Lynch (@michael_irl) 's Twitter Profile Photo

I am a big fan of this “docs for robots” concept of having .md files specifically intended to boost agent performance!!! Super interested to see how this evolves across the ecosystem. What are your thoughts?

Michael Lynch (@michael_irl) 's Twitter Profile Photo

Pro tip: Use ChatGPT as a preprocessor. 1. Describe your vague idea 2. Refine it into exact specs by chatting about it 3. Ask ChatGPT to generate a structured prompt 4. Paste that into your coding agent Better prompt = better results 🚀

Pro tip: Use ChatGPT as a preprocessor.

1. Describe your vague idea
2. Refine it into exact specs by chatting about it
3. Ask ChatGPT to generate a structured prompt
4. Paste that into your coding agent

Better prompt = better results 🚀