With Muse, Dot, and Grokbot all getting so good, itโs hard to justify running and maintaining your own personal agents through Hermes, OpenClaw, a custom Telegram chief of staff.
Keeping mine running and maintaining at this point is mostly sunk cost fallacy.
@Nirmal_Utwani Wow, exactly my thoughts. My pi is not doing great either. Using an SD card as the disk, I burnt one of the cards because I ran my local transcription model on it. My openclaw today is just a bad astra with poor token management running on suboptimal infra connected to an outlet.
Remember these 3 technologies, our world will be reshaped by them:
1. Solid-state batteries: legit chemistry, just waiting on manufacturing yield.
2. mRNA programming: COVID proved it works, we will program cells to fight cancer.
3. Room-temp superconductors: they work today, not yet at usable pressure.
These are premature but correct techs waiting to change the world. Enjoy ๐
Take humans out of the loop.
Your attention and judgement is the most valuable resource. Same goes for your users.
Give clarity and space to AI, and it does the same for you.
Create tasks with the way to validate the outcome. The upfront investment takes you out of the loop.
Profits donโt hallucinate.
You can optimize your agents all you want. Run twenty million agents in parallel on the cloud. Build 80 billion skills and chain them together.
If your business is not making money, you are not creating value. Value materializes as money. If you cannot sell what you ship, stop and change course.
Claude Code Routines are here! In addition to a schedule, you can now trigger templated agents via GitHub event or API โ with our infra & your MCP+repos
They've changed how we do docs, backlog maintenance and more internally at Anthropic
Get started at https://t.co/QIDgEAJrRG
This is actually insane.
Dude hard-coded a WebAssembly (WASM) interpreter into the weights of a transformer, losslessly.
In essence, a computer is running inside a LLM that can actually run computations, not infer or guess a calculation like most do today.
Setting your intention for correctness is a hard and mandatory problem in AI coding. Your code will match your intent, not the solution you think it should produce. Intent is detailed and upfront. Rewriting SQLite in Rust with missing intentions gives a broken database engine. It works but does more harm than good. The same goes for the simplest code you write. The LLM will match your intention, and if there is no way for it to revise and correct itself, you will only get a piece of your intention.
In this example, the intention was to replicate SQLite in C with Rust. It missed key concepts and introduced bugs among 500k+ lines. It is impossible to debug if this is not your passion or job. I imagine more of these bugs creeping into every corporate codebase as we ship faster with limited intent.
If that is the case, how do we convince leadership to create space for engineers to provide intent and iteration cycles for the AI Agents while the shipping speed is lightning fast? ๐ข โก๏ธ
Released today: /loop
/loop is a powerful new way to schedule recurring tasks, for up to 3 days at a time
eg. โ/loop babysit all my PRs. Auto-fix build issues and when comments come in, use a worktree agent to fix themโ
eg. โ/loop every morning use the Slack MCP to give me a summary of top posts I was tagged inโ
Let us know what you think!
Hey @the_yamiteru, from one builder to another, just came here to say that anytime I asked a question like this on X, the responses, positive or negative, did not have any value. However, having observed many successful app/tool launches, if you build something to solve a problem of your own, does not matter the size, just a problem you experience firsthand, and simply build for yourself, only then when you let it out, it has the greatest chance of success and a much higher signal for questions like these. So, build something for yourself, then share it with the world. Chase the problem and the money seems to follow. Donโt get motivated or demotivated by some random person saying โIโd pay for it,โ people say all kinds of things here. Wishing you the best. Keep crafting ๐งฑ
`tabular-nums` should be the default for any number that updates ( timers, counters, prices, percentages, scores, live data etc ).
you can enable this tnum OpenType feature using the CSS property `font-variant-numeric`.
.tabular-nums {
font-variant-numeric: tabular-nums;
}
It's time to bring Haptics to the web ๐ซจ
Create custom tactile patterns with strengths + durations for your web interactions.
Make your app feel as good as it looks โจ
โ https://t.co/DGz9Eu6nto