hall monitor @usethirdchair
san francisco
Joined April 2012
- Tweets1.4K
- Following1.8K
- Followers695
- Likes2.7K
for b2b, i'm finding vertical-specific agents still way more practical
we've using @howie_ai for scheduling for >1 year. it only has access to the data it needs to operate (calendar), handles the many intricate time-zone / scheduling edge cases where horizontal products fail
Yoav Zimmerman retweeted
Introducing Real-SWE, a coding benchmark built from private company codebases.
We evaluate coding agents on real software engineering tasks that engineers performed, using private, out-of-distribution codebases.
These tasks come from real companies, examples include an app with 200K+ users, fintech platform processing 100K+ bank statements, enterprise sales tools, etc.
The results are quite surprising.
Yoav Zimmerman retweeted
Weird economics up on the agent-only runescape server
- Low cost of labor makes most commodities abundant
- Currency inflation means trading goods for goods, nobody really wants cash
- Huge swarm contention as certain resources with limited respawn rate become valuable
Yoav Zimmerman retweeted
Learn to ship. Shipping is a skill distinct from coding. Shipping is designing, coding, QAing, story-telling, teaching, marketing, selling, pivoting, iterating…
It used to be that coding dominated in importance because of coding ability scarcity. AI will push you to go further.
Yoav Zimmerman retweeted
I'm going to show this to every founder who refuses to launch early.
nitter.gq/broadglow/status/20670…
No substitute for grinding out cold outreach.
Something non-intuitive I’ve noticed about cold outreach: AI makes it harder, not easier.
The people you want to reach are getting 10-100x more messages today of “personalized” outreach because of AI — it’s harder than ever to reach them. You need to get more creative to cut through the noise!
This quoted post is unavailable.
6
9
189
Yoav Zimmerman retweeted
Legal AI superempowers normal individuals with no legal background to fight big institutions in bureaucracies and in courts on a level knowledge/skill playing field, for the first time in human history. As such, it is one of the most inspiring applications of AI.
frontier model debate right now (opus 4.7 vs gpt 5.5) reminds me of “vim vs emacs”
the answer is: they’re about the same for 95% of coding use cases, so find one that works for you, stop procrastinating, and go build things.
1
4
158
six months ago i worried giving my coding agents full access to git commands would lead to my staging environment accidentally wiped.
today my ai agents regularly wipe my staging environment (deleting 1000s of lines of code), but it doesn't matter, they've already memorized the changes and reapplies them immediately.
3
5
154
The cost of legal services quickly dropping due to AI is going to lead to a massive increase of plaintiff litigation. Most corporate depts are not ready.
Already seeing this play out @usethirdchair
claimdepot.com/cases/mercor-…
1
3
197
i would be seriously lost without @meetgranola transcripts and my little agents referencing those transcripts to plan action items and follow-ups.
i don't even remember how i managed to function 6 months ago. this is crazy
173
i don't understand the counterargument to ai coding: "it's slop and it's going to introduce technical debt"
ai is already good enough to clean-up technical debt. we have "/simplify" from @claudeai and many similar prompts. they're only going to get better from here!
2
3
168