John Hartquist (@johnhartquist) header
John Hartquist (@johnhartquist) profile picture

swe → mle → aie → swe 🦀🤖

Joined May 2011
Photo by John Hartquist (@johnhartquist)Photo by John Hartquist (@johnhartquist)Photo by John Hartquist (@johnhartquist)Photo by John Hartquist (@johnhartquist)Photo by John Hartquist (@johnhartquist)Photo by John Hartquist (@johnhartquist)Photo by John Hartquist (@johnhartquist)
My first open-source library, a Rust implementation of Alexandre François' Resonate algorithm. - real-time spectral analysis (think STFT/CQT) - arbitrary frequency bins & per-bin time/frequency tradeoff - no windowing or buffering - Python & WASM bindings - live mic browser demo
2
1
181
Hands-free Jevving on the Apple Watch! Inspired by @kunalabcdani's coffee ordering demo, I had to see if I could get something similar working on my watch with "raise to talk" UX. Built with @typesafeai Jev, @DeepgramAI Flux, and my new Rust crate, system2. details ↓
What if we could manipulate UI at the speed of thought? Built this voice coffee ordering system with @typesafeai Jev. The drink and any updates show up while I’m still mid-sentence, ~200 ms from word to screen. Feels pretty magical!
3
2
12
1,462
Handling several items in one sentence uses @allietheicon's shopping cart pattern: ask again, listing what's been found so far, until Jev picks "nothing else". Changes preview while you talk and apply when you pause. docs.typesafe.ai/cookbooks/c…
1
75
With system2, a Rust enum can be the options of a Jev Choice: its doc comments describe them to Jev, and the answer is one of your variants. All three questions go out in one request, and the answers come back typed, ready to decide programmatically.
62
Thought I was already living in the future coding with Claude/Codex/GitHub/Termius on my iPhone/iPad. Then I tried @theo's @t3dotcodes — and wow. Haven't been this impressed in a new dev tool in a long time. Superior UI/UX in almost every dimension. Highly recommend.
1
1
147
New career highlight: having meme retweeted by @typesafeai
we are nothing without our beverages h/t johnhq from our discord
46
John Hartquist retweeted
Also hilarious if anyone thinks I've become a big fan of agent-led Rust out of any other consideration than purely scientific inquiry. Do you know how badly I would hate having to personally write Rust?? But when the premise changes, you need to update your priors.
15
4
1
286
25,025
John Hartquist retweeted
Every developer needs an AI shed: An always-on machine running on their tailscale network where the majority of their herdr agents are running.
623
357
208
7,756
660,964
John Hartquist retweeted
Reject non-understanding. Slop grenades are not just code and PRs. We run a real risk that *reading* gets entirely discounted, because of the exhaustion of getting low-quality and unverified AI prose being thrown at you all the time. There's a thread going viral about a 'perf improvement' attributed to a compiler change. The PR *description* itself has the AI telling this person it's *not* due to the compiler change, it's how the algorithms and data structures have changed. And the implication is that a similar win could have been attained otherwise. I want AI in the service of understanding the universe and enhancing human cognition and creativity.
191
197
52
2,678
146,790
John Hartquist retweeted
This is the the quote I've been citing a lot recently.
you can outsource your thinking but you cannot outsource your understanding
846
4,354
272
46,574
2,785,697
Motivated by training real-time music transcription models, I needed something that integrates with my PyTorch training and runs the same way in the browser via WASM. Rust's portable SIMD auto-vectorizes the hot loop across NEON, AVX2, WASM SIMD128. github.com/jhartquist/resona…
1
59
All credit to Alexandre François for coming up with the algorithm. For anyone interested in digging deeper, his project page has the paper (ICMC 2025 Best Paper), an ADC talk, and the C++ reference implementation (noFFT). alexandrefrancois.org/Resona…
55
John Hartquist retweeted
Have you ever wanted to train LLMs in pure C without 245MB of PyTorch and 107MB of cPython? No? Well now you can! With llm.c: github.com/karpathy/llm.c To start, implements GPT-2 training on CPU/fp32 in only ~1,000 lines of clean code. It compiles and runs instantly, and exactly matches the PyTorch reference implementation. I chose GPT-2 to start because it is the grand-daddy of LLMs, the first time the LLM stack was put together in a recognizably modern form, and with model weights available.
252
1,775
329
12,356
1,662,159
If I had to pick one programming language to use for the rest of my life, it would be Rust. Really enjoying getting up to speed with Rust Web Development by @recvonline 🦀
5
686
John Hartquist retweeted
Introducing AI Brainstore! It’s a proof-of-concept of a brain for an AI agent. Ask an agent a question, it checks its memories for an answer, otherwise it browses the web and learns the answer. As it learns, it saves memories to its brain. GitHub: github.com/mckaywrigley/ai-b…
124
207
36
1,634
514,748
John Hartquist retweeted
Nobody is better positioned to make a great living in the AI/ML space than software developers. If you know how to build software, you have some of the most valuable skills necessary to make an impact.
10
30
356
62,544