Tweets about AI, finance and startups Building @tryatomiq
New York, USA
Joined June 2025
- Tweets594
- Following604
- Followers755
- Likes2K
Pinned Tweet
Atomiq + Jev is looking speedy ⚡️
We're making this feature available to everyone on tryatomiq.com very soon.
This is going to be a short lived phenomenon.
I bet that all of these companies will come out with their own subagents if they haven’t already.
These subagents will be able to connect to whatever primary orchestrator you are using. But there will be clean ownership / handoffs.
Of course companies should not just hand over their API to Claude directly.
The importance of having a product that AI interfaces with efficiently cannot be overstated. It will be a huge part of a good user experience.
Are Claude or GPT going to optimize specifically for exceptional UX while interfacing with Delta airlines? Probably not.
But if you’re Delta - that’s pretty important. Makes total sense to control the inference around their products.
notice..
> Delta is blocking AI agents from booking flights
> JPMorgan Chase and Charles Schwab block agents from accessing accounts
> Yelp requires paid data licensing for agents
> Amazon has blocked Grok bot and other shopping agents.
the internet is starting to block AI agents that were supposed to do everything for us
I’m so bored of this take.
Prove that Opus 5.5 can generate consistently growing revenue for a business - without human ideation or intervention.
That’s gonna be pretty tough when your competition is also… Opus 5.5
I guess no one is reading the T&Cs and privacy policy on this Cosign thing?
As written it seems there's nothing preventing a16z from sitting around in a partner meeting and reviewing this 'private signal' data that other VCs submit.
Maybe I'm missing something but I don't typically email my competitors every time I see an opportunity I'm interested in.
One of my favorite features of @cosign is the private signal feature.
LinkedIn connections are hard to parse. An X follow can mean a lot of things. A DM is high friction.
On Cosign you can passively signal private intent: would meet, would fund, would work with, etc, and get introduced if there's mutual interest.
Paying for Fable and Astra feels a lot like paying for lawyers.
There’s a lot of work being done, I’m being billed for all of it and I don’t know if I need any of this.
This will cure any one of their AI pessimism. The future is bright.
Through the VOICE trial, Terry is using his Neuralink implant to help fine-tune a brain-to-voice interface for himself and others who can’t speak. He trained the algorithm first by miming speech as best he could, then by simply thinking the words and hearing them come out in his own natural voice.
Powered by Grok Voice from @spacexai
Simple guide to Jev:
A) deterministic (use code)
B) probabilistic with bounded outcomes (use Jev)
C) probabilistic with unbounded outcomes (use an LLM)
There was vibe coding, now there’s vibes as code
Cool idea here
I've seen people describe Jev as an "AI if statement". But what if it actually WAS an if statement?
Introducing Probably: a programming language powered by Jev: probably-lang.southpolesteve…
Jev baked into the language. “feels” asks a question. “match” routes between descriptions. “while” keeps going until something stops feeling true.
This is obviously a toy, but it's fun to think about what something like Jev unlocks. Jev makes the decisions, an LLM does the writing, and a little program ties it together.
So brutal...
Similar situation for me yesterday, I just asked gemini to copy a document into 3 different folders and make some changes to each.
It couldn't do it. Walked me through manual copy / paste in google drive.
“We’re building prod not god”
Great tagline
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI?
I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev
• 20-200x faster
• 40-400x cheaper (w/ output tokens free)
• Frontier composable intelligence optimized for decisions
AFAICT the shortest path to AI-based economic revolution
We have gone from people prompting agents to agents prompting people
This is one of the biggest barriers to adoption for consumer agents.
Most everyday people look at a blank text box and have no idea what they can even ask AI to do. Agents should be proactively suggesting tasks based on your context.
So well articulated from @gdb 👇
