CTO @testudoglobal. Insurance for AI ✝️

NY
Joined November 2018
Throughout my entire career I don't think I've ever opened the GitHub notifications page
2
150
A great prompt for this: “Create an html artifact with every possible edge case handled and verify everything looks solid visually. Interview me with any questions about questionable cases.”
One of my recent favorite AI use cases: asking it to try and break the stuff I built. Adding lots of data, long names, unusual emails, labels, etc. Coming up with worst case scenario basically. This is a vibe-coded demo, but you get the idea.
3
770
Give Claude a rule that every single visual change needs to open an HTML preview with before & after screenshots. Thank me later.
1
7
451
PSA for who needs to hear it: Peptides are not considered safe yet
2
7
1,313
I think figma went public at a great time, but damn I just don’t see a need for design software anymore.
3
339
Nothing resets your motivation like SF. I can spend weeks in insurance feeling like we’re lightyears ahead, then come back here for 8 hours and feel like we’re not moving fast enough.
3
24
1,380
Alec Dewitz retweeted
Replying to @beknabdik
just switch to extra usage
1
6
1,514
Sign of a good design engineer: Fixing the layout shift in stock shadcn login/signup forms caused by password managers like lastpass
7
1,305
Fun fact I used to try to sneak in design fixes like this in my PRs when working in big tech, but frequently was told by management that it was out of scope or my team “didn’t own it”, so was forced to undo my changes.
In today's episode of "why is it aligned like that?" I present to you the Microsoft Teams chatbox and info banner lockup
7
1,775
If you do nicotine daily, pivot to not doing nicotine
1
19
1,546
Alec Dewitz retweeted
Recap from my hyrox in NYC with @alecdewitz
1
8
724
This will mark the beginning of responsible and efficient AI use within orgs. The wild west of AI tinkering on the company dime is slowly coming to an end. From what I’ve heard from employees at big tech cos, companies are spending upwards of $500M per month on tokens alone.
NEW: Uber is reportedly capping employee use of AI vibe-coding tools at $1,500 per month after blowing through its AI budget.
3
502
My experience this past week using GPT 5.4 has been that Codex is strong as a standalone agent, but not very effective as an orchestration layer for Claude agents. Claude can spin up multiple agents on demand, and I've found its coordination and aggregation of their outputs is much clearer for data driven tasks.
4
418
Another week, another AI engineer meetup with agent insights! London builders favorite coding agents included: - Claude Code (58%) - Codex (14.1%) - Cursor (11%) - Antigravity (3%) - Other: OpenCode, LangChain(?), no preference (13.9%) Pretty consistent numbers over a 2-week period and no major flippenings!
A lot of alpha last night from the AI engineers meetup in NYC. Among the 79 attendees, favorite coding agents included: - Claude Code (54.4%) - Codex (11.4%) - Cursor (8.9%) - Antigravity (3.8%) - OpenCode (2.5%) - Other: Gemini CLI, Amp, Roo Code, Composer, Ur Mom (19%) All with an overwhelming favor towards Opus 4.6, and Codex 5.2 coming in close 2nd.
3
2
4
901
One line of code just cut our GitHub Actions from ~10 min to ~4 min
You really should stop using Github's action runners. I just made a 1 line of code change and cut our CI times by 50%.
1
3
431
At the cafe today and this 15 year old kid next to me was building a reminders app for his entrepreneurship class project. Not Lovable. Cursor + Claude Code Love to see it
1
6
443
A lot of alpha last night from the AI engineers meetup in NYC. Among the 79 attendees, favorite coding agents included: - Claude Code (54.4%) - Codex (11.4%) - Cursor (8.9%) - Antigravity (3.8%) - OpenCode (2.5%) - Other: Gemini CLI, Amp, Roo Code, Composer, Ur Mom (19%) All with an overwhelming favor towards Opus 4.6, and Codex 5.2 coming in close 2nd.
2
2
4
26
4,458
The single highest leverage thing you can do for your health is wake up at sunrise and go for a walk with your friends.
1
2
275
Quite interesting that current models consider scrolljacking a "luxurious" web experience
3
262