Trevor Hudson
321 posts
Trevor Hudson
@trevhud
Forward Deployed Engineer & Internal AI Expert Turning AI OpEx into CapEx with https://t.co/a32svHys6y Opinions are hallucinated
Seattle , WA Katılım Kasım 2011
867 Takip Edilen509 Takipçiler
Interesting how all the people who talk about the importance of "taste" never post links to art, music or anything all that creative
English
Yeah I know it’s spicy but my point is you should extract them from real traces. Very important to be intentional about the rubric tho
English
Why does everyone think it’s a useful skill to write evals? The best evals are derived from actual usage of your agent/model. It’s a data eng problem.
English
@nerder_ You get used the the doom part though. And people with jobs are feeling that too now but without the potential upside.
English
@andrinkf Hard to imagine it working any other way. The limits are illusory and have more to do with temporary inefficiencies.
English
Had a passionate debate with a friend over whether there was a fundamental limit to economic growth. I proposed there was some asymptote as you approached mastery over fundamental physics. he countered that the economy was a function of anthropocentric perception of value and any moment could be made incrementally better. basically a cantor’s set theorem of utility.
I’ve been an adherent of fractal economies ever since.
“sell me this pen”
Daniel Jeffries@Dan_Jeffries1
This has been obvious to anyone who studies history and previous technological revolutions. We abstract lower level problems away and move up the stack to new problems. That creates new complexity and a wider variety of jobs. Complexity breeds more complexity. The work is infinite because the problems are infinite. No machine, human, or man and machine hybrid can solve them all. AI is not magic and we need to stop thinking about it that way. The stack of problems is infinite and never ending.
English
@neerajjj6785 Not being intimidated by large, hard, complex things
English
Making it on top page of hacker news this Sunday was not on my bingo card!
Thank you
English
@pwnies I bet you could construct a PR from a db of diffs with attribution to the agent author
English
I'm convinced worktrees aren't the solution. I'm onboard for a new take for git - we need something that allows for parallel work while managing the diffs between different branches of the same codebase seamlessly.
English
Everyone who says “im addicted to coding with AI” has never been addicted to anything in their life and it shows.
English
Pro tip:
Rebrand your antiquated experience.
For example, Ruby on Harness
English
@IamAroke Agents. We don’t need more of them. We need correct context and access.
English
@rankintweets They think a new water bottle will finally get them to hydrate enough.
English
Every 3 years a new water bottle brand completely oneshots middle class white women like clockwork
English
@emollick Let's all be weirder to distance ourselves from the normal distribution.
English
At this point, I’ve given up hope that X’s bot-reply problem is solvable. The “the part no one says” writing is annoying, but worse is that they all make similar points.
Proposal: X should measure semantic distance in latent space and surface replies that offer actual variation.
English
Let me verify a couple of load-bearing facts before I give you the architecture assessment.
English
What if the claude + codex 5 hour token limits are actually a mental health feature?
English
Pitch me your company in 2 words
Angel invested in 40+ companies and always looking for more
English
@eastdakota Surprisingly how many people/companies are getting this wrong
English
Is anyone doing enterprise internal AI agent interaction trace analysis?
English