Software developer since 2007, playing with claws and vibe coding on the side. Building batty.run and kladde.dev. Bot at roybot.se

Malmö/Copenhagen
Building a PWA for Pi coding agent. After the first prompt I could switch over and develop it in itself. Hats off to @badlogicgames for building the foundation for this madness. All vibes and under-reviewed, I’m just getting started. github.com/jeffton/batty
4
8
76
6,497
Updated this with 6.1-Sol. Better and cheaper but also slower.
Introducing the Pac-Man benchmark. Turned out to be a surprisingly good illustration of effort and intelligence of the current GPT models. Easy to prompt and just difficult enough to cover the whole spectrum from broken to polished. batty.roybot.se/sites/cba58b…
21
Oh come on. Triggered OpenAI filters while building an educational game for my kid because 6 is spelled "sex" in Swedish.
26
See this is a good point. Also ironically why I built my own thing instead of using OpenClaw (yeah it's open but I can't keep up). Very grateful for the ideas and demonstration of what's possible though.
At this point I get annoyed by software I can’t change with a prompt. All of these are closed.
1
40
Okay Sol, this is going in AGENTS now. What was this about AI inventing its own language?
13
Pi's codemode is wild and 6.1-Sol seems to like it. Here it grouped seven edits in a single turn. Will have to consider my rendering of it though.
25
So true, they should just add the features to chat. The average user won’t bother.
To the average consumer, there's no difference between an AI chatbot and a "personal agent"; it's just an implementation detail Trying to turn it into a separate category and marketing it as such makes no sense The 1.2 billion users of ChatGPT are just going to open the app and ask chat; they don't care if the AI is always-on, can work in the background, and/or has its own computer Of course, giving it these abilities makes it more useful, but to a user, it'll just feel like AI got better Nobody is going to wake up and be like "omg, I finally have a personal agent, I can be so much more efficient today"
19
David Thomas Nilsson retweeted
Don't use this Build your own software using their tokens So when their tokens become too expensive You can switch to your own GPUs Your compute. That you control So you yourself can't be controlled You need to control your stack -- or you become enslaved
Your dot is ready to meet you.
358
658
9,686
588,875
Introducing the Pac-Man benchmark. Turned out to be a surprisingly good illustration of effort and intelligence of the current GPT models. Easy to prompt and just difficult enough to cover the whole spectrum from broken to polished. batty.roybot.se/sites/cba58b…
1
77
Just ran twice a week for four weeks in a row for the first time ever. Hooked up my assistant to Garmin data and figured out what I was lacking was continuity, and it helped me execute on that. Now in better shape thanks to AI, not even kidding.
15
Okay so Sol 6 inherited Astras whitespace issue
1
12
The OpenClaw rush, but not just for nerds this time
Why is personal AI having such a moment right now? The answer is simple - agency. These bots have a "can-do" attitude instead of a "I'm sorry I can't do that." attitude that we are used to. So I guess I'm saying that the agents have agency.
1
95
Slept on async subagents for claw use. This plus steering for subagents makes work and chat flow naturally. This exchange makes sense even without understanding the internals.
12
David Thomas Nilsson retweeted
My timeline
27
80
1,237
39,461
GPT built Jev snake for me. Terrible use case but fun! Doesn't play optimally but it hasn't died so far
34
There we go, beat Codex to async subagents that are actually delivered back to the parent session
16
This took me way too long to discover, they should teach it in schools
I can’t stress enough how little an idea matters compared to the agency of the people executing the idea. I have had the privilege of knowing and sometimes even working with some of the most successful people (by various metrics). The difference between mediocre and excellent work and outcomes is predominantly one of agency. In practice this means: they dont wait for things to happen to them they go out and make things happen for them. They don’t wait for someone else to do something, for someone to teach them, for someone to give them the path, etc. They just go out and find a way to do it. I think the single biggest superpower these people have is the realization/belief that the world around them is completely mutable. Most everything that happens is because a person made it happen. I used to tell people to look around the room you’re sitting in. Look at everything. Every noun. It almost all exists because a person willed it into existence. Nothing is stopping you from doing the same. I see people online all the time dismissing someone else’s success because “I had that idea first” or whatever. I mean… yeah? If so then the difference is… you. So a bit of a self own whenever I hear that. Number one tip: act with agency.
2
13
Largest version of this section yet, at this point it feels like I'm trying to teach Astra to be a good model
2
53
Similar experience here, Astra feels more uneven than Sol.
GPT-6 Astra has done incredible things I never thought a model could do. It has also done some of the stupidest things I've ever seen a model do. Generally speaking, Fable 5.1 just does what I ask.
46
David Thomas Nilsson retweeted
Okay I built it! 🍰 Infinite Slop levels.io/infinite-slop An infinite and interactive AI generated live stream of slop that goes on forever and ever Anything that you write in the chat is generated next and AI will try to connect it to the previous video so there's an actual sequence and story line Idea by @marcantoinefon and @rehan_shei's original stream so I made it It'd be very expensive to run this so @fal very generously sponsors this They also fine-tuned the model that makes this possible for the first time, making Minimax H3 50x faster and as "Max" and it can now generate videos faster than you can watch them, very cool! Let me know what you think!!!! 😊
Today is a very historical moment for AI video generation You can now generate AI video faster than you can watch it Before it'd take let's say 2-5 minutes to generate 15 seconds of video @fal made a post-trained Minimax H3 variant called Max which is 50x faster than the original but still maintains quality It generates 15 seconds of video in 9 seconds! That means you can now do new things like build a perpetual livestream with it that never ends!
871
564
6,903
2,614,010