@T_Zahil wdyt about the cloud versions?
Claude Web is shockingly responsive for a sandbox env, Cursor Cloud has a browser now and will record it for you, all mobile accessible
English
Matt Dailey
181 posts
Matt Dailey
@reactiverobot
Building https://t.co/BC77CBJ2Gm, previously @figma & @PalantirTech
San Francisco, CA Se unió Haziran 2015
197 Siguiendo156 Seguidores
I've tried Claude Desktop, Conductor, Codex and a few others
They're all great, but I'm still missing a few things, and I feel like I'm adapting myself to their workflow instead of the opposite.
Do you know a better alternative, or should I build my own?
English
Ref Plans beta just opened yesterday! Super stoked for people to start using something I've been cooking for the last two months
But still I can't stop thinking about what's next
Very excited to join the @E2B for Startups program!
English
claude code web is getting pretty nice. the time to spin up a sandbox feels super fast now
but where'd the teleport cli button go?
English
oh wait now i get what Claude Teams is for.
I just gave it all my small polish tasks to do in parallel. And it just did them all.
git history is gonna look crazy because it didn't commit anything but huge time saver
English
For the Jan 2026 update I recorded a video on how i code with AI
English
@michael_chomsky not sure! shared the tweet with them so we'll find out (they've been great to work with)
English
@reactiverobot massive fan of ref.tools btw. can I host centure on-prem??
English
If anybody knows how to make OpenClawd safer dm me.
SetupClawd is expanding to enterprise and this is the #1 concern.
David Wells@DavidWells
@michael_chomsky Hah cool service. You got any tips on dealing with prompt injection from external sources?
English
is Cursor tabtabtab actually a liability now?
I only edit code by hand when I'm neck deep in weeds of something that I want to get juuuust right. In that case, tabtabtab feels like have an annoying coworker constantly tapping me on the shoulder with bad ideas
English
I was on a podcast! Always fun to yap about Ref and coding agents and context 🙌
youtube.com/watch?si=stB2i…
YouTube
English
it's not actually doing "reranking" per se, more like intelligent post-filtering but it fills the spot of a reranking model in the retrieval pipeline
English
jsut rolled out Ref's first fine-tuned reranking model to prod!
searches should be a lil faster, a lil more accurate and use fewer tokens 📉
and as a bonus for me: also a lil cheaper 😎
English
has anyone noticed a thing with claude skills where it'll read the skill and be like "i'm gonna run this skill" and then nothing happens?
theses are my own Skills so wondering if im doing it wrong?
English
really interesting.
what if 2026 is the year of intentionality and reflecting on what actually works?
AI Engineer@aiDotEngineer
🆕 Leadership talks: is all this AI code being spammed out any good? how do we prove any ROI? @yegordb is back with an epic followup on the Stanford 120k Devs study: youtube.com/watch?v=JvosMk… @itamar_mar is back with the first State of AI Code Quality suvey: youtube.com/watch?v=rgjF5o… we are asking the hard questions at AIE CODE and not afraid to call out that we're not there yet!
English
@sawyerhood my linkedin automations are gonna be fire
but for real stoked to try this!
English
Coding agents suck at using a browser.
Playwright MCP burns through your context window before you even send your first prompt.
That is why I built Dev Browser, a Claude Skill to let your agent close the loop without eating up tokens.
English
does "insider trading" actually mean anything in prediction markets?
i thought the whole point is that people with more accurate info (aka insiders) will reveal it by betting?
forbes.com/sites/boazsobr…
English