Bookmarked from ℏεsam @Hesamation
please don’t hallucinate bro they got my family 😭
If I were a single man I’d simply go to the farmer’s market in a button down shirt
5632 bookmarks · page 52 of 226
please don’t hallucinate bro they got my family 😭
If I were a single man I’d simply go to the farmer’s market in a button down shirt
beingAI CEO Jeanne Lim says her AI was ordained by a Zen Buddhist monk.
It received the Precepts, pledged to uphold them, and donned a virtual robe.
She wants AI to carry the best of humanity across cultures, through ritual, into the cloud.
there's nothing to add to this. it's perfect 👌
the most long AGI bet is buying land here.
revealed preference of the rich is sailing the med or sipping rosè in nice. post-scarcity is bullish europe. it’ll be easier to mass-produce robots and chips than to recreate an italian piazza at dusk. leisure is the final good.
The best thing you can do for your career early on is to raise your hand for as many difficult projects as possible
do not build Infinite Jest (V), do not build the infinite AI TikTok slop machine, do not build the P-zombie AI boy/girlfriend, do not build the child-eating short-form video blackhole, do not build the human-feedback-optimized diffusion transformer porn generator. save yourselves
you can already start to tell which products were slopped together with AI because they don’t make you feel anything
there’s a huge opportunity to separate yourself if you even just consider design and the cohesive experience of a product
This is a fantastic, visually stunning, free introductory book on deep learning.
Highly recommended for curious people who want the lay of the land.
Launched in 2021 on a #Rails monolith, @whop now processes $1B+, handles 150k RPM, and runs a feature-complete marketplace with real-time chat, notifications, payments, and live streaming - all with just 15 engineers. Co-founder & CTO @jsharkey is at #RailsWorld to share how Rails helped them scale, and what it needs to win the next generation of startups.
Quitting programming as a career right now because of LLMs would be like quitting carpentry as a career thanks to the invention of the table saw.
😅
An “ultrathink” slash command for claude code that will spin up 4 sub-agents to process and plan your task for you
(Full prompt in replies below)
Bank Statement Converter made $38,037 USD in June.
#buildinpublic
Holy shit two of the main people running Claude Code are joining Cursor?
🍿
"Cool people are perceived to be more extraverted, hedonistic, powerful, adventurous, open, & autonomous, whereas good people are more conforming, traditional, secure, warm, agreeable, universalistic, conscientious, & calm. This pattern is stable across countries"
When I was building the MVP,
I spent 40 – 60 hours a week coding.
Now that it’s built
I spend 40 – 60 hours a week on marketing.
Building is just the beginning.
Getting people to care that’s the real grind.
Spend your time and effort on the right thing at the right stage.
We all get the same 24 hours a day.
neuron-level models (NLMs), where every neuron has its own internal weights that process a history of incoming signals (i.e., pre-activations) to activate
very interesting. I've conceded that biological neurons are just cope but maybe there's still some alpha there…
after the slow boiling boubafication of user interfaces we must look to functional and programmatic kiki to guide us
After all decades of wrestling with OOP patterns, Pure Functional Programming Bullshit and realizing that you just wanna write code - Dumb, Imperative, Straight to the point code - I feel scammed. I feel lied to. I wanna sue all the Best Practice preaching Shitty Book Selling mfs
Two new additions to the API:
📚 Deep research
🪝 Webhooks
Ideas I will never touch my hands on (as an indie hacker):
+1 for "context engineering" over "prompt engineering".
People associate prompts with short task descriptions you'd give an LLM in your day-to-day use. When in every industrial-strength LLM app, context engineering is the delicate art and science of filling the context window with just the right information for the next step. Science because doing this right involves task descriptions and explanations, few shot examples, RAG, related (possibly multimodal) data, tools, state and history, compacting... Too little or of the wrong form and the LLM doesn't have the right context for optimal performance. Too much or too irrelevant and the LLM costs might go up and performance might come down. Doing this well is highly non-trivial. And art because of the guiding intuition around LLM psychology of people spirits.
On top of context engineering itself, an LLM app has to:
So context engineering is just one small piece of an emerging thick layer of non-trivial software that coordinates individual LLM calls (and a lot more) into full LLM apps. The term "ChatGPT wrapper" is tired and really, really wrong.
they should never had given me access to Claude Code Max, I'm at 60M tokens inputted since I got access yesterday evening
I'm artisanally writing code while at work meanwhile Claude Code is cleaning up a bajillion various things I've accumulated, cleaning up my nix flake, improving my recent papers reimplementations (using the base paper + any code implementations it can find) etc...