Co-founder & CEO @19thlabs → You can just train models.
Joined May 2010
- Tweets 2,954
- Following 1,069
- Followers 873
- Likes 3,177
316 Photos and videos
Jun 8
It’s becoming increasingly common for our roadmap discussions to become “should this be a feature or a skill”
2
7
648
Jun 3
Subjective observations:
1. Codex does better work at standard speed
2. Opus 4.6 is still the best writing collaborator
90
May 29
Domain pricing is so opaque. We paid $4k for 19th.com, whereas an obscure 6-character name with no search volume the owner wanted $500k
May 29
Just bought our .com for $25k from a public NYSE-listed company, cold-calling its CEO.
> realize the domain name belonged to a public company (shit).
> cold-call all board members then CEO
> follow-up everyday
2
279
May 5
Myself and @spencerthurstn started @19thLabs.
19th is a tool that lets anyone build a model in <60m. Adding them to the list of things "you can just build".
We're opening the beta to more folks soon. Just signing up lets you experience my favorite detail on the launch site.
4
3
26
2,403
May 5
At Uber we both saw 1 MLE often spread across 3 PM's. All having to fight to get one of their 20 ML experiments prioritized that quarter.
19th lets more folks build models. In the same way Codex lets more people code. The codex moment for ML?
1
3
235
May 5
We’re also hiring two founding engineers.
ML experience is optional, ML curiosity required.
DMs are open.
2
6
248
Feb 26
Thesis: The way we think about capturing product feedback is constrained by the way we had to process it.
Excited for a future of free-form voice note feedback from users that gets cleanly categorized, summarized, and "emotion scored" by LLMs
1
224
Feb 4
Very nice - we played with something similar recently, specifically for icons. Runs locally in a fun lil Mac app or via CLI.
Imagine turning any image into a vector (now you actually can)
→ Convert raster images into editable vectors
→ Simplify and control color output
3
1
23
7,269
Feb 4
My favorite detail from the release site is this preview of the path quality coming out of the model.
1
2
182
Feb 1
I do think we'll see the rise of "authenticity" as a value pillar alongside quality.
The woodworker vs Ikea.
Already happening in UGC and a lot of graphic design.
Feb 1
I wonder if there's a market for *actually beautiful & intent-fully designed simple apps.
Been browsing the App Store for some really simple apps (think Weather, Water counter etc.) and it's INSANE how ugly most apps are.
Pure AI slop trash but... they have great reviews for some reason.
Most people are probably satisfied with "it gets the job done" but there must be people like me who kinda identify themselves with the types of apps they use?
I don't want to use those ugly apps. I want to "feel great" when I use them.
Stupid example maybe but there's people that use the cheapest $20 suitcase they can find and there's people that want to use a $1000 Rimowa.
I think despite all the AI slop out there, there's a chance to stick out with good design.
The challenge is just getting your app surfaced in the sea of the slop.
I think people like @asallen are a great example of that idea.
Small shower thought, but I'm kinda tempted to try a few things.
2
394
Feb 1
What timeframe is required for a prediction market to become a slot machine?
1
147
Jan 27
We're cranking on a lil model that can draw.
Cute, endearing, and slightly brain-breaking to see its progression through epochs match that of a child learning to draw.
1
3
195
Jan 26
Resurfacing with some new things soon. But also, gaussian splats have been an interesting way to augment training data. Especially good for drowning out perspective from why a model learns two things go together.
1
144
New year, same mission — keep the world adventurous forever.
📷 @MaxDiCapua @currysizzle
9
19
313
16,798