Dominic Elm @elmd_

Founding Engineer @boltdotnew • Lead for AI/ML and WebContainer • JS • Node.js • Rust 🦀 • WebAssembly • Deep Learning 🤖 • Amateur Athlete • Ironman • Dad learn-wasm.dev null Joined October 2014

Tweets

21K
Followers

8K
Following

980
Likes

36K

Vite ⚡ @vite_js

2 days ago

🚀 Vite 8.1 is out ⚡ Experimental Full Bundle Mode: ~15x faster dev server startup & ~10x faster full reloads on large apps 🧩 Chunk Import Map to keep chunks stable & improve caching 🦀 WASM ESM integration 💡 Lightning CSS on the road to default 👀 and more! Blog post below.

24 227 2K 148K 300

View Details

Cline @cline

3 days ago

We've kept hearing how GLM-5.2 beats Opus 4.8, and are skeptical of benchmarks - so we tested them on a real bug from the Cline repo. While both models fixed the issue, GLM was the winner in terms of cost and code quality: - GLM used twice as many tokens (GLM 1.1m vs Opus 660K) but cost half as much (GLM $0.41 vs Opus $0.81) - Opus finished quicker - 1.6 min and 12 tool calls vs GLM 4.7 min and 28 tool calls - GLM cleaned up dead code and verified the build compiled before completing. Opus didn't - it left type errors that passed tests but broke the production build. Both runs used the same Cline harness prompting and tools, so it seems GLM is RL trained to spend more tokens verifying its work before completing. Impressive work by the @Zai_org team!

219 609 8K 867K 2K

View Details

Dominic Elm @elmd_

5 days ago

@burkov But what about the limited context window? You have to compact much more often and also it’s still not multimodal.

0 0 0 418 0

View Details

Dominic Elm @elmd_

a week ago

I don't think I'd frame Applied AI as "just plumbing" or simply plugging models together. There's definitely a difference between classic ML engineering and Applied AI, but the best Applied AI work still has a lot of depth. This includes building evals, improving prompts systematically, fine-tuning, RL, smaller task-specific models, reliability work, data science, latency/cost tradeoffs, and a lot more. I'd argue that a lot of the hard problems are in making AI systems actually work in messy real-world conditions. That's exactly the kind of work we're hiring for.

0 0 0 250 0

View Details

Dominic Elm @elmd_

a week ago

We're hiring a Senior and Staff Applied AI Engineer @boltdotnew ⚡️ Come build AI agents that turn natural language into production-ready apps. Remote, global, no degree required. Timezone must predominantly overlap with US work hours. Links in 🧵

16 26 200 17K 143

View Details

Dominic Elm @elmd_

a week ago

@Zigdexk34 Hey! Relevant industry experience is required. But definitely check out our careers page.

1 0 1 222 0

View Details

Dominic Elm @elmd_

a week ago

@dakshgup Looks very cool. Congrats!

0 0 0 487 0

View Details

Dominic Elm @elmd_

a week ago

@jkomyno @every Yes very true. It’s quite token hungry but I noticed that it was able to find really deeply rooted issues that other models weren’t able to find. It’s always a tradeoff and I wouldn’t default to Fable for everything.

0 0 1 14 0

View Details

Dominic Elm @elmd_

2 weeks ago

Not gonna lie, but I miss Fable 5. Not for vibe coding but it's such a strong model to find deeply rooted issues that other models can't find. Here was my workflow: Plan (Fable) -> GPT 5.5 (build) -> Review (Fable)

1 0 4 383 0

View Details

Dominic Elm @elmd_

a week ago

@poistudioltd Got me! Nice one.

0 0 0 7 0

View Details

Dominic Elm @elmd_

a week ago

This is why Linear feels so incredibly good to use. There's a lot of details they got right. This level of details matter linear.app/now/invisible-…

0 0 4 472 5

View Details

Dominic Elm @elmd_

a week ago

Love the details in the top one. Nicely done and looks smooth.

Ali Grids @AliGrids

2 weeks ago

The best way to polish CSS animations: slow them down to 0.1x v/ @hugosaintemarie

8 19 579 55K 472

0 0 2 671 2

View Details

Dominic Elm @elmd_

2 weeks ago

You cannot satisfy an LLM. It will always either find something or hallucinate. My god, I just finished 6 /goal loops and it’s still finding gaps. Will there ever be an end? Only one way to find out 🥸

1 0 4 525 0

View Details

Dominic Elm @elmd_

2 weeks ago

@kuizinas @Meligy I know right?! Let’s gooo!

0 0 1 31 0

View Details

Gajus @kuizinas

2 weeks ago

zod-compiler compiles Zod schemas into zero-overhead validation functions at build time. This makes Zod validation 2-74x faster. github.com/gajus/zod-comp…

21 30 469 33K 213

View Details

Kimi Developers @KimiDevs

2 weeks ago

🌘Kimi-K2.7-Code Weights & code We explore endless possibilities together with the community🤝 huggingface.co/moonshotai/Kim…

45 226 3K 132K 453

View Details

Anthropic @AnthropicAI

2 weeks ago

The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Claude models is not affected. We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible. Read our full statement: anthropic.com/news/fable-myt…