So maybe the lesson isn’t:
“These companies are just wrappers.”
Maybe the lesson is:
A great wrapper + distribution + positioning can itself become an enormous business.
As the founder put it:
“Sab marketing hai, boss.”
I asked a founder:
“How does a voice-to-text app get valued at $2 BILLION?”
He smiled and said:
“You’re looking at the technology.
Investors are looking at something else.”
That conversation changed how I look at AI startups 🧵
The code might be reproducible.
The model might be replaceable.
The underlying capability might even become free.
But distribution?
Habit?
Brand?
Polish?
Being the first product someone thinks of when they say “AI dictation”?
Much harder.
Starting September 14, we're permanently raising standard weekly limits in Claude Code by 25% for Pro, Max, Team, and seat-based Enterprise plans. Until then, the current 50% increase will be in place.
Starting September 14, we're permanently raising standard weekly limits in Claude Code by 25% for Pro, Max, Team, and seat-based Enterprise plans. Until then, the current 50% increase will be in place.
Okay I built it!
🍰 Infinite Slop
https://t.co/wz0a7VSBtS
An infinite and interactive AI generated live stream of slop that goes on forever and ever
Anything that you write in the chat is generated next and AI will try to connect it to the previous video so there's an actual sequence and story line
Idea by @marcantoinefon and @rehan_shei's original stream so I made it
It'd be very expensive to run this so @fal very generously sponsors this
They also fine-tuned the model that makes this possible for the first time, making Minimax H3 50x faster and as "Max" and it can now generate videos faster than you can watch them, very cool!
Let me know what you think!!!! 😊
AI video generating faster than playback changes what TV can be.
Viewers could create scenes, interact with the cast, and steer a live reality show as it unfolds.
You can test this at https://t.co/GUS3PMccEq
You can create a 5 seconds it's super fast.
Minimax H3 Max has generates video faster than you can watch it so I hooked it to a twitch livestream! Now you can watch infinite interdimensional cable - link to the stream below
Watching the current rush to build everything with AI has made me pause.
AI has changed how quickly we can create software, but it hasn’t changed the questions we should ask before building it: Do we actually need it, and will we still use it once the excitement wears off?
But here's the problem: 69% of developers report more bugs in production when using AI-generated code. Meanwhile, 58% worry about the security risks.
Speed up now, fix later. Banks can't afford "later."
Here's the real problem: LLMs do okay with text tasks, but they really struggle with image, video, or website generation. This matters because more tools are pushing AI-generated UI builders and dynamic content. The models aren't ready for that complexity.
Open-source models did even worse. 65% accuracy. The gap between proprietary and open-source tools is massive. If you're relying on open-source LLMs for code generation in critical paths, you're not getting the reliability you might expect from a 2026 system.