solving my own problems

Localhost
Joined June 2023
Voise has officially joined the @cartesia Startups Grant Program! We're building real-time voice AI into our product with Cartesia. #BuildingWithCartesia
1
1
715
reminder to use `scrollbar-gutter: stable` if your layout shifts when the scrollbar disappears
28
84
6
1,956
86,824
Zée retweeted
Design Engineering 101: Make your number steppers feel awesome. A number stepper that only bumps once per click makes you click twenty times. Reward the hold. Animate it nicely with rolling digits, blur on transitions and gradient masks to avoid abrupt cuts.
20
36
8
908
74,985
i thought of the same thing. but if you're deep into UI's and interfaces, you'll find they increased the UX far way more on these new landing page of theirs.
the new @vercel website looks so generic, its sad. the old one was a gold standard for developer branding. what happened?
3
372
oh shit. i found my aha moment!
139
I really miss the days I debugged manually
1
2
194
I just messaged some random people on the in Internet. Not in a selling manner but mentorship I guess. Well, let's wait
2
5
200
Hello @WhatsApp this emojis might turn out to be the worst feature yet y'all have shipped. I almost reated with the laughing emoji on a friend's parent burial.
1
4
450
just published the first version project and it's a production reliability runtime for AI agents. npm i agentsist
Finished the Princeton paper on agent reliability. Found myself relearning math I haven't touched in years just to understand their metrics. Confirmed this is the thing.. Building it because nobody else is.
1
3
258
as AI agents move more into production, the winning infrastructure won't be the framework that helps agents think. It will be the runtime that helps agents survive.
1
116
Finished the Princeton paper on agent reliability. Found myself relearning math I haven't touched in years just to understand their metrics. Confirmed this is the thing.. Building it because nobody else is.
Found a Princeton paper published this February on AI agent reliability. Turns out nobody is solving this properly yet — the field optimizes for smarter models, not more reliable ones. Confirmed what I've been building for 19 days has a real name and real research behind it.
1
1
478
Found a Princeton paper published this February on AI agent reliability. Turns out nobody is solving this properly yet — the field optimizes for smarter models, not more reliable ones. Confirmed what I've been building for 19 days has a real name and real research behind it.
Didn't write much code today. Spent the entire day on LinkedIn, Reddit, and X researching competitors, use cases, and whether people actually have the problems I'm solving. Turns out they do. Tomorrow I find the first developer struggling with exactly this and I reach out.
1
3
434
Didn't write much code today. Spent the entire day on LinkedIn, Reddit, and X researching competitors, use cases, and whether people actually have the problems I'm solving. Turns out they do. Tomorrow I find the first developer struggling with exactly this and I reach out.
Day 17, I fucked up with typescript types
1
1
355
Day 17, I fucked up with typescript types
Still had some issues today waiting for npm to reply to me
1
1
275
Still had some issues today waiting for npm to reply to me
my sdk on the ai-sdk
1
320
the suffering is real!
2
3
251
my sdk on the ai-sdk
Day 15 building an AI start-up from Africa. I built the sdk and it works very fine(well on localhost). Will be finalizing the publishing tomorrow!!
1
1
282
Day 15 building an AI start-up from Africa. I built the sdk and it works very fine(well on localhost). Will be finalizing the publishing tomorrow!!
Day 14 building an AI start-up from Africa. Thought on the building of the sdk
1
2
333