Zack Chapple

Faster, cheaper, and higher quality! Who says you can't have it all. Coming Soon 👀

@zeeg It's not that they used open source, its that they violated the license on an open source model that has a modification for commercial products above a customer count and revenue. And even then its an attribution clause which would have been easy to add.

I cant believe Cursor would allegedly use open source. Dont they know they're supposed to be stealing everyones data instead? This cannot stand!

Built a small dice probability tracker with @CloudflareDev Workers and Cron Triggers. Every minute, a worker rolls 2d6, writes the result to D1, and updates a live tally board so you can watch the distribution take shape over time. dice-app.shane-58d.workers.dev

@zeeg @KentonVarda The highest I got myself was just over 1k a day, but again I'm an outlier too.

@Zackary_Chapple @KentonVarda i myself cannot spend more than 300-400 a day unless im pushing it through heavy token consumption automated tools (and those are often wildly inefficient spend) and im far above the average person in volume of code

I think I have only used like $1000-$2000 of AI tokens total in my life so far. I guess Jensen would fire me.

@zeeg @KentonVarda depends on level of parallelism, reverse engineering. To be fair I may be friends with a few of the outliers in the world.

@Zackary_Chapple @KentonVarda i mean this in the nicest way possible, but i call bullshit the highest impact engineers i know would struggle to extract $100k/year spend as value

@zeeg @KentonVarda Disagree, I know one or two folks doing 50k+ a month and making significant impact and doing entire years worth of work in a month.

@Zackary_Chapple @KentonVarda The only way most folks are spending 250k is through waste. That’s gas town level phony research.

Preventing death by a thousand paper cuts

This is why I love strong enterprise leaders using our platform. One we're working with saved 9 months on an initiative, rather than cutting the team they moved forward their roadmap because they actually have one.

@zeeg @KentonVarda A 500k engineer at nvidia is expected to lead and help forge the best path for ai usage and consumption. That is not something a few thousand in token usage a day will do.

@KentonVarda I wish folks like him would stop making ridiculous statements like this. Everyone latches on them non stop.

I believe even for you it’s a missed opportunity. Not because you need ai to be successful, obviously not, but because as you bend ai to your will the knowledge gained in how you do it is more valuable to the rest of the organization. I use @kdy1dev as an example. Bro has no need of ai to be brilliant. However learning how he uses ai and makes it do what suits him not only increased his velocity but benefited the rest of the team.

MiMo-V2-Pro & Omni & TTS is out. Our first full-stack model family built truly for the Agent era. I call this a quiet ambush — not because we planned it, but because the shift from Chat to Agent paradigm happened so fast, even we barely believed it. Somewhere in between was a process that was thrilling, painful, and fascinating all at once. The 1T base model started training months ago. The original goal was long-context reasoning efficiency. Hybrid Attention carries real innovation, without overreaching — and it turns out to be exactly the right foundation for the Agent era. 1M context window. MTP inference for ultra-low latency and cost. These architectural decisions weren't trendy. They were a structural advantage we built before we needed it. What changed everything was experiencing a complex agentic scaffold — what I'd call orchestrated Context — for the first time. I was shocked on day one. I tried to convince the team to use it. That didn't work. So I gave a hard mandate: anyone on MiMo Team with fewer than 100 conversations tomorrow can quit. It worked. Once the team's imagination was ignited by what agentic systems could do, that imagination converted directly into research velocity. People ask why we move so fast. I saw it firsthand building DeepSeek R1. My honest summary: — Backbone and Infra research has long cycles. You need strategic conviction a year before it pays off. — Posttrain agility is a different muscle: product intuition driving evaluation, iteration cycles compressed, paradigm shifts caught early. — And the constant: curiosity, sharp technical instinct, decisive execution, full commitment — and something that's easy to underestimate: a genuine love for the world you're building for. We will open-source — when the models are stable enough to deserve it. From Beijing, very late, not quite awake.

@jonathan_wilke @sseraphini @nextjs Time to switch to something based on Tan stack

Holy crap... with the @nextjs 16.2 update the next server is now using 10GB+ of memory and my MacBooks fan is running constantly (before it never even turned on). What's going on here...

kimi is here. workers ai is officially in the big model serving game blog.cloudflare.com/workers-ai-lar…

Actual magic!

this is magic

@_mchenco @threepointone On launchpad cohort announcement day too!!

today

Not a slop fork

@WSJ Can confirm it works wonders!

Eli Lilly’s experimental weight-loss drug retatrutide met the primary and key secondary endpoints in a trial, showing significant reductions in weight loss and blood sugar levels. on.wsj.com/4uSOB7Y

@GergelyOrosz Heard some interesting things from our customers about similar things, and 3-5x increase cost of the plans

I am hearing tons of complaints from Cursor customers at enterprise companies: A silent change put almost all models Cursor uses behind Max mode. Devs who used to manage to “spread out” monthly credits over a month see all of it used up in 1-2 days. Are furious + switching.

MiMo-V2-Pro and MiMo-V2-Omni are now free in OpenCode many of you have tried them in stealth - Hunter Alpha and Healer Alpha. Mystery solved worth mentioning - MiMo-V2-Pro: - ~1T params - 1M context - optimized for coding