Umami bytecode

Joined September 2009
Replying to @misamajic
I OVK se raduje
1
1
12
112
new insult: "you would be easy to distill"
1
17
1,182
US Amazon £410 + import ~£600 UK Amazon £962.50
140
I guess this is a new take on open source
We keep saying we want open-source frontier agents. Fine. Then let’s build the dataset. @badlogicgames, creator of Pi, just shared some of his agent traces used to build Pi on @huggingface. I’m now sharing some of mine too, exporting them from @hermes, @opencode, and Claude via @tracesdotcom, and I’ll keep going. Why this matters: one of the biggest bottlenecks for open-source agent models is the data. And all of us are generating that data every day through our conversations with agents. If enough builders share even a slice of their traces publicly, we can create the largest crowdsourced open dataset for agents. Time to put your tokens where your mouth is and give a chance for open source to win!
136
i'm about to put the most embarassing data on the internet. hope i don't get pwned. you can do so too! updated pi-share-hf, share your coding agent sessions with the world, savely (fingers crossed, read the README.md in full) github.com/badlogic/pi-share…
11
9
150
11,878
Apple dev program signup as a company. They want that "professional" look. A few minutes in LLMs and you can generate all of it. Why? Because I had an intentionally spartan looking website, based on which they rejected me. Will revert, as soon as approved.
117
AI didn't kill software engineering It killed the illusion that shipping code and engineering a system were the same thing
32
53
13
462
23,830
Love @whimsical for sketching things and remember thinking that it's a shame that it'll become irrelevant in this age of LLMs. But little did I know that they've been working on their own MCP that allows you to export your diagrams as Mermaid markup. Very nice! Thank you
98
dotemacs retweeted
this email could have been a 70-turn conversation between our agents with 100 tool calls.
19
105
3
1,683
58,970
Getting AI slop via email: okay. Getting AI slop from people I had personal relationships with for years via email: man that hurts. Stop doing this! Please stop. I want to read your typos and bad English over this.
23
8
2
308
13,738
This can be summed up as: When using a tool, use the best practices. Especially when using it with an agent.
If reading this kind of thing gives you a nasty stress response, know that "Thankfully, they helped me restore the database, and the full recovery took about 24 hours."
153
What model + agent are they using in Kuwait?
169
Does anybody stop & think how the mercenaries feel? They’ll no longer be able to invade a country without a plan.md clearly specced out.
2
108
dotemacs retweeted
my favorite thing on this site is all the life and productivity advice i get from all the random web devs who think they have made it
3
1
30
958
Went to an online meeting with other devs. Lots of old assumptions, approaches and even some that are refusing to use agents on ethical grounds.
1
114
🇨🇳❤️
51
flicker company: "guus, you cool.with us training on all your GH repos, books, etc. right? it's not infringement, it's transformative work!" also flicker company: "How dare you pay us and then train on our precious model outputs! HOW DARE YOU! It's against our ToS!"
We’ve identified industrial-scale distillation attacks on our models by DeepSeek, Moonshot AI, and MiniMax. These labs created over 24,000 fraudulent accounts and generated over 16 million exchanges with Claude, extracting its capabilities to train and improve their own models.
12
21
2
355
12,900
"no one will vibe code their own software"
Replying to @claudeai
So the winners of the Claude Code hackathon were: - a personal injury attorney - an interventional cardiologist - an electronic musician - an infrastructure/roads systems worker - and one software engineer That should tell you something.
22
1
2
97
24,864