Kev

@bcardarella @th3mus1cman @elixirlang Haha alright man, my bad

@KevConti @th3mus1cman @elixirlang But you jumped into my thread, I don't have to get your points. That's the point.

Something else about AI generated @elixirlang - it *LOVES* to generate code with dynamically created atoms, thus exhausting the memory on machine. Just putting in a rule to always use strings isn't the right solution either for perf. I haven't yet found a good way to prevent OOM issues like this without Boomer Coding the fix.

@bcardarella @th3mus1cman @elixirlang I guess to me it just doesn’t matter if AI were to mess this up on its first edit, because the outcome is that it gets it right after it sees the Credo warning. To me, it seems like you are missing that point

@th3mus1cman @KevConti @elixirlang It's amazing that people are missing the point. Anybody can cobble together a toolset to handle this. The issue is does AI generate good Elixir or not, as people keep insisting. It doesn't.

@TSR_Jonah My wife says Craig Anderson

The first person to bang the drum at a Sabres playoff game needs to have a few specific qualities. It needs to be someone who: a) Will get a huge reaction b) Hasn’t been seen in awhile c) Represents the drought as a whole In my opinion, that person is Jeff Skinner.

On what might be the final day of the Sabres playoff drought, one last look...

3. This takes four roles - A product role - A senior architect role - An engineer role - An AI enablement role And still scales with number of codebases.

3. AI Capability Engineer It's a full time job to make the AI good at working in your codebase. This person works on integrating w/infra, teaching it to query all data, and continually addressing failure cases in the codebase.

I think this is close/kinda right. Studies show 80% of features deliver net-0 value to the end-user. AI doesn't fix this, if anything it exacerbates it. I think we need to do 3 things:

I'm not the only one doing this. - karpathy best thought leader, best person to learn from imo. Nanochat is the best way to get into training LLMs its the simplest and most digestible source for building your first AI model - steipete This guys GitHub is a national treasure, his writing is also very strong. Peekaboo, summarize.sh, openclaw, oracle, just talk to it, etc.. all unique and very useful - badlogicgames Mario’s Pi is a staple AI engine and possibly the best, simplest, open source agentic loop to learn from. Despite what people say about his methods, I think he’s going to set some new standards for Open source contribution. Big respect. - TheAhmadOsman This man is the GPU king, giveaways and lots of dense educational content around self hosting and home inference. He’s also tight with pretty much all the open weight labs and has them on for interviews regularly - sudoingX This is an up and comer who will change the game, he's pushing the limits of what a single gpu can do - Ex0byt I can confidently say this man will be fundamental in making local inference on massive models possible. - alexinexxx I genuinely feel motivated by her drive. She’s a real hard worker learning about GPU kernel programming. Also good aesthetics - gospaceport I would not have gotten into building my own hardware without this man’s hard work. He’s taught me so much about hardware and the economics of this. He also has the most impressive homelabs I’ve ever seen. - alexocheema The founder of Exolabs, pioneering Apple hardware inference, he’s also very engaged in the community and a good guy all around. If you are interested in Mac minis and Mac Studios this is your guys. - nummanali This guy is so prolific, he’s made tons of CLI tools for managing llm subscription budgets, using Claude code with alternative models etc.. - thdxr The entire Opencode team is wonderful but Dax specifically is a good writer. More anti-doomer content to sooth your anxieties. - juliarturc If you are interested in the science, Julias channel is where it’s at. Almost everything I’ve learned about LLM compression has been from her. - Teknium The Nous research & Prime intellect teams are both some of the most hard-working and principled people around. Tough fight in an industry so aggressive. - victormustar Head of Product for Huggingface, enabling us all to publish our work. - louszbd Head of community at ZAI some of the top LLMs available right now that are open weights. They supercharged the movement - SkylerMiao7 Making frontier intelligence fit on 10k USD of hardware. Via MiniMax - crystalsssup Building the best Open Weight model on the market, and releasing their latest research before their next gen model. Believe it or not these people are carrying the entire industry and giving us a fighting chance.

@0xSero @OpenAI support Open AI ^^^

Putting out a wish to the universe. I need more compute, if I can get more I will make sure every machine from a small phone to a bootstrapped RTX 3090 node can run frontier intelligence fast with minimal intelligence loss. I have hit page 2 of huggingface, released 3 model family compressions and got GLM-4.7 on a MacBook huggingface.co/0xsero My beast just isn’t enough and I already spent 2k usd on renting GPUs on top of credits provided by Prime intellect and Hotaisle. ——— If you believe in what I do help me get this to Nvidia, maybe they will bless me with the pewter to keep making local AI more accessible 🙏

Composable agent flows in Elixir. Mix deterministic workflows (FSM) with LLM orchestrators — they nest freely at any depth. Human approval gates, parallel branches, and checkpoint/resume built in. Built with Jido (thanks @mikehostetler!) #ElixirLang github.com/lostbean/jido_…

@OpenAI @sama are these quantized versions of GPT-5.4 or a different approach?

GPT-5.4 mini is available today in ChatGPT, Codex, and the API. Optimized for coding, computer use, multimodal understanding, and subagents. And it’s 2x faster than GPT-5 mini. openai.com/index/introduc…

God I love Elixir. We just don't have this problem that is killing other languages. Work with an immutable, non-nullable language for your AI projects. Just do it. #MyElixirStatus

@dan_note super cool! is there support for fuzzing/obscuring PII by element for privacy/compliance or is it a near-term priority? that'll be my big blocker to replacing Microsoft Clarity

New package! PhoenixReplay — session recording and replay for Phoenix LiveView. → Records assigns server-side, replays by re-rendering the actual view — pixel-perfect, not a DOM approximation → Zero client-side JS — no bundle size impact, invisible to users → See actual server state during replay: changesets, Ecto structs, form data → Navigation, page transitions, live_patch — all captured in one session automatically → 30s session = 8 KB rrweb and friends record DOM mutations client-side: they fight CORS, break on shadow DOM, drift when assets change, and produce approximate replays with visual glitches. None of that applies here — LiveView templates are pure functions, so same assigns = same HTML. The BEAM just keeps the state. More coming soon. github.com/dannote/phoeni…

If your Claude Desktop was hanging over the past 24-ish hours, this is why.

i think that the Sabres got better at the deadline this year. not drastically better, but they should be more resilient in extremely physical games, which was a weakness they had. final grade, 6 out of 10

overall, i was still hoping to see i see one more move at the forward position. since that didn't happen, i'm even more confused why we paid extra to get salary retention on Schenn.

okay last #sabres post for me now that the trade deadline is over... (thanks for being patient with me AI friends)