AI Lead, Red Hatter, automates clouds, former memory leaker, enjoys all things AI Engineering, Containers, Ansible, Python. Climbs, MTBs, trail runs, kayaks.
Don't get me wrong, big fan and early adopter of #MCP from late last year. But at the same time MCP is the biggest gift to hackers since C Memory Management.
Every PM should be using Claude Code.
So I built a HUGE course for you to learn Claude Code... IN Claude Code!
🔹 Complete guide
🔹 Make PRDs, analyze data, create decks
Soon, I'll sell it for $149.
For the next 24h: FREE!
Follow + RT + comment "CC" & I'll DM it.
@karpathy I've been monitoring mine and in certain hotels they are fabulous. I actually have a severe hearing loss so not noise related for me. Factors I suspect: great curtains, generally great beds with tons of space, great cooling - 18C for me. Also I'm frequently tired on the road
@browsercompany much as I love(d) Arc the hours of fighting it for simple SSO logins means it is back to Safari - Arc is now failing 20, 30, 40 times a day. An exercise in frustration (cleared cookies, multiple reboots etc etc)
Heading to New York for the AI Engineering Summit with a big focus on AI Agents tomorrow. The event for AI Engineers
https://t.co/4D8IOUaJHb
#AI#aie#AgenticAI
So temperature has always been a dial I've used crudely, both in AI programming and hotel rooms - making big changes. This article certainly helped me get a better understanding of the why, and what exactly happens when you tweak the dial #AI https://t.co/mxHvwbV0nq
@sabre1041 Leaving Las Vegas, just don't watch it in a room full of spirits (alcoholic spirits that is) Perhaps you could recreate it with your favorite carbonated drink and match Nicholas Cage drink for drink (maybe go low caffeine?)
For quick, short, easy edits I'm in neovim but I'm "between IDEs" right now, Python, Ansible, markdown, asciidoc. So VSCode, Windsurf, or Cursor (or is there something else?) I really would like clean access to both remote and local LLMs
@sabre1041 Yes I know your pain, now I travel less that's one pain point reduced. I try and do them on the go on my mobile which is fine on credit card centric trips
Tired of using deepseek-r1 (yes i know the ollama/lm-studio are not "true" deepseek). Mistral's latest. mistral-small-24b is out. If you are on 32GB then choose your quantization wisely. No regrets on my 128GB :) https://t.co/vdsKDNAi0V
There goes the day... I was planning on working through some more PydanticAI material, and I decided on my way to browse the HuggingFace smolagents. I'm quite intrigued; both are very clean and Pythonic. I must write a few beyond hello worlds to test smolagents... #ai#AgenticAI
System, and Human, Prompts are rapidly becoming the new stored procedures, which didn't end well for those .Com database migrations.
How long before the new refrain is
"Works with my LLM"
#AI#promptengineering#AIAgent
@sabre1041 This would be a good benchmark activity for OpenAI's new Operator or AGI. I don't think I've ever succeeded in improving Lufthansa seating when booked on a United ticket. We should build a suite of "frustrating challenges" and benchmark agentic frameworks against them.