yashad

"wait so how are you actually using AI in your daily life?" ongoing thread, chasing curiosity 🧵

@sathaxe @KristyT Want one!

After two years of fatherhood and two kids I can confidently say that these are the best dad pants of all time if you're a SF dad with a kid 0-4 I'll buy you a pair to prove it

@Shmall This is dope, I would love to try, how do?

They say your phone knows more about you than your mom. So why can't Siri tell me things like how much I spent on food delivery this month? DM me if you want to try a phone that can.

@levie @tszzl Lmao the definitive statement, I love this. Spending more time and energy on things that you just wouldn’t/couldn’t have done before which is Good! Nominate this to be the AI optimist mantra.

@tszzl I’m here using AI to automate something that would have taken me 10 hours before to do, so I wouldn’t have done it; and now I’m spending 2 hours on all the follow up work. So basically 2 more hours of work than before. The doomers can’t comprehend this.

tbh I only feel more accelerationist as the capabilities ramp … the scaredest I was was in early 2023

@saranormous @msg Yes please!

any nontechnical folks want to get more comfortable/powerful in their use of AI and want to be a beta user on something I made?

@itsiddharth_ awesome

some collage and texture studies 💫 finally finished this guys

Some real time moments from oryzo.ai

@igarakii whooooooa

motion design reel from last year #motiondesign #aftereffects

@kylebrussell Sweet, which model did you get?

@yashad The idea is that it's going to be like a fully customized Alexa that also has cameras and moves in an animated way so it'll be a fun character that can see and hear Over time will switch to powering it with local models running on a Mac mini/Mac Studio in the closet

Ordered my daughter her first robot, a Reachy Mini, for her sixth birthday 🥳

@johnpalmer Whoooa this is it!

New logo for Mesh, a laser company.

@metapreston 100%

OpenClaw is Linux. We need a MacOS for AI

@localghost Is the context window limiting at all?

5.3-codex-spark is insanely fast at responding on openclaw. not as friendly but actually a big experience step up

Yeah! Reluctant to give it too many permissions but slowly opening up with security audits, love the ability to switch models per task but retain core memory/context + setting it up to be future proof so my Obsidian vault = the brain, all state and memory stored there, agent is just execution layer. If something comes along that’s better, I can just point that new thing to my vault and everything can be rebuilt as it’s clearly documented. Reminds me of the idea of being able to take your social graph everywhere via decentralized social or fediverse.

Bought a new Mac mini to properly tinker with claws over the weekend. The apple store person told me they are selling like hotcakes and everyone is confused :) I'm definitely a bit sus'd to run OpenClaw specifically - giving my private data/keys to 400K lines of vibe coded monster that is being actively attacked at scale is not very appealing at all. Already seeing reports of exposed instances, RCE vulnerabilities, supply chain poisoning, malicious or compromised skills in the registry, it feels like a complete wild west and a security nightmare. But I do love the concept and I think that just like LLM agents were a new layer on top of LLMs, Claws are now a new layer on top of LLM agents, taking the orchestration, scheduling, context, tool calls and a kind of persistence to a next level. Looking around, and given that the high level idea is clear, there are a lot of smaller Claws starting to pop out. For example, on a quick skim NanoClaw looks really interesting in that the core engine is ~4000 lines of code (fits into both my head and that of AI agents, so it feels manageable, auditable, flexible, etc.) and runs everything in containers by default. I also love their approach to configurability - it's not done via config files it's done via skills! For example, /add-telegram instructs your AI agent how to modify the actual code to integrate Telegram. I haven't come across this yet and it slightly blew my mind earlier today as a new, AI-enabled approach to preventing config mess and if-then-else monsters. Basically - the implied new meta is to write the most maximally forkable repo and then have skills that fork it into any desired more exotic configuration. Very cool. Anyway there are many others - e.g. nanobot, zeroclaw, ironclaw, picoclaw (lol @ prefixes). There are also cloud-hosted alternatives but tbh I don't love these because it feels much harder to tinker with. In particular, local setup allows easy connection to home automation gadgets on the local network. And I don't know, there is something aesthetically pleasing about there being a physical device 'possessed' by a little ghost of a personal digital house elf. Not 100% sure what my setup ends up looking like just yet but Claws are an awesome, exciting new layer of the AI stack.

@danielpearson @ycombinator @wideframeai Congrats! Would love to try!

Wideframe just launched on @ycombinator! Do hours of video work in minutes with @wideframeai. Check out our launch: ycombinator.com/launches/PUI-w…

Yeah. Maybe it’s my style of learning but I often reflect on how public school, through its various ways (perhaps well intentioned) made me dislike the idea of learning and being curious because the environment just wasn’t conducive to it and it wasn’t until college or shortly after that I managed to crawl my way out of that mindset, a few excellent professors in community college helped!

I do ~10 minutes a night of math with Lu (age 5) and he is so hungry for this stuff that 2nd grade would be a bore to him now. He is amply ahead with barely any instruction. My most salient memory of school was not from learning, it was the waiting

@mikesmith187 isleofany.com doing some cool stuff!

A friend of mine is looking for the best marketing agency. I'm not hyper tied into that world but who's doing great work out there right now? Mother crushed it with Anthropic, so something of that caliber would be nice.

still finding it hard to definitively know which model is the best one for what i need it to do: not more, not less. and the tradeoffs associated - whether it be speed or cost or context window. feels like were either underpaying or overpaying. maybe the reason that clarity doesn’t exist is that it’s fundamentally misaligned with the business model of the labs but it would be hugely beneficial for consumers (me)

@whatdotcd Yoooo this looks slick, I wanna try!

I LOVE SOULSEEK. I HATE REKORDBOX. So i made a bunch of new updates to SoFTTSeek to take me closer to removing Rekordbox from my life. *Search for Music/Videos *Library that imports from and auto syncs to Rekordbox *Play and sort Audio & Videos w/ Subtitles *Sick Visualizers Also works for Intel & Mac Chips now. Reply if you want it and I got you.

@simonsarris Yeah! It’s always explicit wholehearted praise for seamless execution and empathetic collaborative realignment for mistakes. IMO a crucial way of communication to retain as we flip from (potentially) talking to more agents than humans in day to day life.

I think you should say 'please' and 'thank you' to LLMs, because it is communication, and all communication is practice