CEO @zoodotdev, 👩🏻‍💻 @oxidecomputer, 📝 @ACMQueue

Falken's Maze
After two years, fewer than 30 engineers, and a fraction of the capital poured into earlier efforts, we are officially launching v1 of Zoo Design Studio, this includes our own CAD application, CAD kernel, AND ML model! Please give us feedback! zoo.dev/blog/zoo-design-stud…
44
64
644
184,051
The same idiocy that did this is probably going to reinvent firmware, replace it w markdown “boot computer, make no mistakes, figure it out, oh it needs WiFi too”
Replying to @jessfraz
Even this MD skill per device seems a bit insane and just like burning tokens to me, you are taking something deterministic and making it non-deterministic github.com/facebookincubator… But I guess this is the future, just not the one I’ll take part in sorry seems just wasteful
6
3
98
12,920
I’d suggest using home assistant since it has support for 99% of everything and if you don’t like the support just patch it versus hand rolling your own interface Personally I use home assistant and then the home assistant APIs to pipe into custom interfaces
Andrés Karpathy is using OpenClaw based AI agent to take care of his home automation. To autonomously discover smart-home devices, learn their interfaces, and wire them into a unified dashboard and command system. Instead of manually using separate apps for Sonos, lights, and other smart devices, he built an AI agent that can control everything for him. He just types normal instructions, and the agent figures out the technical steps needed to make things happen. --- From @NoPriorsPod YT channel (link in comment)
8
2
73
8,264
Even this MD skill per device seems a bit insane and just like burning tokens to me, you are taking something deterministic and making it non-deterministic github.com/facebookincubator… But I guess this is the future, just not the one I’ll take part in sorry seems just wasteful
5
4
66
15,351
I honestly was shocked when I saw this was how it was implemented, I mean they could reinvented their own home assistant interface with real code and that would have been less wasteful than this
21
2,861
Jessie Frazelle retweeted
I'm starting a podcast, "Confession," where I invite tech veterans to share their painful lessons. We all connect through shared pain. Hear founders & CEOs like Satya Nadella (Microsoft), Dara Khosrowshahi (Uber), Luana Lopes Lara (Kalshi), and others share their stories and lessons learned. Subscribe to Confession on... ✅ Spotify: open.spotify.com/show/6x2J7r… ✅ Youtube: piped.video/watch?v=sd5spHYX… ✅ Apple: podcasts.apple.com/us/podcas… @SatyaNadella @Dkhos @luanalopeslara
69
79
408
49,418
if it was really super intelligence it would have said you are crazy to fly into lax, make it burbank
dots demo. Now... with better WiFi.
6
2
140
9,289
it actually makes me sad everyone treats the semianalyis people like gods when thats what Charlie Demerjian has deserved for long before those people came around
2
17
1,289
Is united giving status transfers?
CEO of @Delta Trashes @elonmusk To Explain Rejecting Starlink — “It’s All True What They Say About Him” viewfromthewing.com/delta-ce…
5
4
141
43,526
Congrats to SpaceX on the launch! Exciting times!!
2
1
51
4,584
I think about this post every time I see someone on here bragging about a perf improvement w agents. “The people who blindly trust agent output are in the former camp. They're sheeple, overdrinking from a fountain of mediocrity.” I love AI too but you need to think!
I've got an agent in a loop optimizing a renderer with the goal to minimize frame times (and tests to measure). It got times down from 88ms to 2ms and allocations down from ~150K to 500. Sounds good, right? Wrong. This is exactly why agent psychosis is a big fucking problem. As an experiment, I rewrote the Ghostty core render state in Go, with access to identically laid out data structures as Ghostty and the exact same validation tests. I made a purposely naive renderer (simple, correct, but slow). 88ms per frame with 150,000 allocations (horrendous, lol)! I then kickstarted a Ralph loop to bring the frame times down. I told it it can't modify input data structures or the public API or tests (they're correct), but it can do anything else it wants. It got to work. It has worked for about 4 hours. I've spent around $350 on this experiment so far. The results? 88ms => 1.5ms 150K allocs => ~500 allocs Incredible right? Nope. My hand-written renderer I ported has frame times (same benchmark) of ~20us (0.020ms) and 0 allocations in the update path. This is the problem with psychosis and lacking systems understanding. If you don't understand the system, you're going to accept that this is an incredible result. If you understand the system, you'll see better solutions immediately and can do roughly 75x better on throughput. The people who blindly trust agent output are in the former camp. They're sheeple, overdrinking from a fountain of mediocrity. Standard disclaimer: I use AI all the time. I like AI. The point I'm making is to not blindly accept results. Think. Analyze. Learn.
28
29
533
83,534
Honestly, glad that happened. It’s a good way to figure out who still has their brain on or not
We are aware that codex is down and are working hard to bring back normal service.
6
1
57
12,774
Actual conversation I had with spectrum today this is like weaponized incompetence
Made with AI
15
3
301
21,540
It’s weird to me Amazon would do this when a purchase is a purchase. They’re making money either way kinda of fucking stupid
51
3
120
44,246
I have a cli for making phone calls if you want that for your codex / agent whatever github.com/jessfraz/switchbo… docs are here github.com/jessfraz/switchbo… it got me a refund on some vitamins that melted in the mail the other day
20
6
236
20,454
As designing hardware gets easier and easier the bottleneck becomes manufacturing Talked to someone on the bus today who said this is the busiest he’s ever seen IMTS, it’s happening
12
4
73
9,182
My phones are also always naked, def skills issue
I've never had a case on any iPhone, I've never cracked a screen or broken the phone. People used to say "oh wait until you have kids." Well, I have a couple now, and one is 3. No problems, clearly a skill issue. No cases for me, using the device as it was intended by design.
17
104
30,767
Going to IMTS in Chicago next week, can’t wait to see all the machines! If you are also going reach out!
6
1
21
4,943
Now more than ever, you can just build things and man, it’s fun
5
9
108
4,859
Part 2: Width Matters, ImpedanceGate and other review, this is fun :) blog.jessfraz.com/post/home-…
It's been over 5 years since I wrote a blog post, but I was motivated. Here's how I am making custom hardware for my house using codex because fuck control4 and all those other fuckers. blog.jessfraz.com/post/home-…
1
1
16
5,527