Dan Adler
@DanielNealAdler
Ops/finance/data and a lot of coding @sourcegraph @ampcode. #T1D.
I keep hearing "shipping fast is the only moat" But having taste is SO much rarer
In 1st- and 2nd-gen AI coding tools you now see a proliferation of modes, toggles, and buttons. This feels off. AI UX done well should mean we don't have to point/click or get lost in a sea of rules or modalities. Sure, you may still have to type or articulate your intent, but…
Been waiting for this one! Context window tips :chefskiss: In 6 months, will we even be talking about this anymore? Will the best agents just seamlessly and invisibly delegate to subagents/oracles? @AmpCode
In this episode, Beyang and Thorsten discuss strategies for effective agentic coding, including the 101 of how it's different from coding with chat LLMs, the key constraint of the context window, how and where subagents can help, and the new oracle subagent which combines…
It's just better
I’m not one for AI hype but @AmpCode is the real deal. Tried it today after listening to @thorstenball on the Changelog podcast and it is hard to describe how good it is. I’ve never been this excited to build
A pretty big moment. The first time my conversations really, ACTUALLY feel better from having multiple LLMs involved. @ampcode just lets them do what they do best.
Amp has a new tool
Fun fact: I found out my account had a bug preventing me from getting FREE credits. The Amp team fixed it within 30 mins on Saturday night and sent me even more credits. I'm talking 8:30pm on Saturday -- and I wasn't even a paying customer.
Every dev tools sales rep and founder knows this argument well! But as a user, there's just never been anything that so obviously pays back multiples as this gen of @AmpCode @claude_code etc., right? How can these cost so little
WILD people think I'm spending "so much" on AI eng tools, when this is my YTD eng spend on @chatprd: - claude $91 - @v0 $120 - @cursor_ai $335 - @ChatGPTapp $1500 - @DevinAI $2500 - @zapier $3621 - human engineers $100k Dumb not to make that 100k as efficient as possible.
All we think about is how to make Amp the very best. Not the best that fits into $__/month, not the best that uses only models from a particular model provider, but the very best.
Still rocking @AmpCode. Not sure why I can't get into Claude Code, maybe I'm just not a CLI guy, but Amp is the IDE-embedded equivalent (same Sonnet model anyway). I love how it can gobble up ALL the context needed to do complex tasks *really* well. Like, what I would do myself.
I just can't get over much this job has forever changed @AmpCode
I can't promise that Amp is right for every company today, but I can commit with everything I have that if you get Amp, it will stay at the model<->product frontier and won't become obsolete. That's why we went multitenant-only, no model selector, consumption pricing, etc.