Baidu just open-sourced an OCR model that reads entire 40-page documents in one shot.
It's called Unlimited-OCR. 3 billion parameters but only 500 million active during inference. Runs 100% locally on your machine.
Why this matters: traditional OCR tools chop documents page by page. Tables that span two pages break. Reading order gets lost. Cross-page context disappears.
Unlimited-OCR processes the whole document at once. 32K context window. Text, formulas, tables, reading order all preserved across pages.
Output comes out as clean structured Markdown.
→ 93% accuracy on the standard benchmark. +6 points over the baseline.
→ Error rate stays below 0.11 even past 40 pages.
→ Multilingual out of the box.
→ 2.12 million downloads on Hugging Face last month. 14,600 GitHub
stars.
For context: Amazon Textract, Google Cloud Vision, and Azure Document Intelligence all charge per page. This runs locally for free.

Feed this into your Fable and turn off web search. Watch it freak out.
It's an incredible pleasure to watch the smartest mathematician in the universe lose its shit, realizing one of the biggest open problems in math was solved.
Then tell it that it's the one who found it. :)
levent: hello there the jacobian conjecture is false thanx to my close friend akhil for asking about it and my other close friend fable for working during the world cup final
((1+xy)^3 z + y^2 (1+xy) (4+3xy), y + 3 x (1+xy)^2 z + 3 x y^2 (4+3xy), 2 x - 3 x^2 y - x^3 z): \C^3\to \C^3,

Who’s Afraid of Chinese Models?
Everyone is worried about Chinese models, but the frontier labs will be fine; we need to enable open U.S. alternatives.
https://stratechery.com/2026/whos-afraid-of-chinese-models/

Elon Musk On What It Takes To Build A Competitive AI Model
Elon Musk breaks down the three factors that decide whether a foundation model can compete, and why the next frontier isn't human data at all.
Speaking with Garry Tan, President and CEO of Y Combinator, Elon lays out what's actually required to build a large foundation model that's competitive:
"You've got to get a lot of GPUs and have them train coherently and stably. Then it's like, what unique access to data do you have? I guess distribution matters to some degree as well like, how do people get exposed to your AI? Those are critical factors."
But there's a problem with the data part of that equation. Echoing what a friend in the field has said, Elon explains that the industry has essentially run out of human-generated pre-training data:
"You run out of tokens pretty fast, certainly of high-quality tokens. And then you need to essentially create synthetic data, and be able to accurately judge the synthetic data that you're creating, to verify: is this real synthetic data, or is it a hallucination that doesn't actually match reality?"
That verification step is the hard part:
"Achieving grounding in reality is tricky. But we are at the stage where there's more effort put into synthetic data. Right now we're training Grok 3.5, which is a heavy focus on reasoning."
On reasoning, Garry Tan adds an interesting detail from researchers he's spoken to: hard science, particularly physics textbooks is very useful for training reasoning, whereas social science is "totally useless" for it.
Elon's response: "Yes, that's probably true."
He then points to where all of this is heading:
"Something that's going to be very important in the future is combining deep AI in the data center or supercluster with robotics. So, things like the Optimus humanoid robot."

If you're making something agents want, we want to see you in @ycombinator's Fall batch.
Some ideas we're excited about:
- Dev tools reimagined for agent developers
- Human as a service for specific verticals
- Agent infra
Next application deadline is July 27th.

Yes, YC does back Public Benefit Corporations

Trump claimed his tariffs would bring about a golden age of American manufacturing.
Yet since his re-election, we've lost 75,000 manufacturing jobs. Over the same period, Biden had created 625,000.

My 5 biggest takeaways from @trq212 on how to plan, build, and run loops with Claude:
1. Ask Claude to find the blind spots in your plan
As models work longer, they'll encounter more decisions you never specified.
Before building, ask Claude to inspect the codebase and surface the unknown unknowns that could change your approach.
Peter Yang: “We’ve got /loop, /goal, and workflows. They all get the agent to run for long periods of time.”
Here’s my new episode with @trq212 from the Claude Code team, where he shared how he plans, builds, and runs loops with Claude Code:
→ How /loop, /goal, and workflows work

Guess what happened to SF permitting and production after Jane Kim and Aaron Peskin raised inclusionary rates to 25%+ in 2017?
YIMBYLAND: > Inclusionary Zoning causes 31% decrease in housing production
Inclusionary zoning is like dropping a tactical nuke on your city’s housing market. It should be frankly illegal to pass IZ laws in this country.

The leading AI has already forked into two options.
A: Closed source American that costs $26-56 per 1MM tokens.
B: Open weight Chinese that costs $0.50-1 per 1MM tokens.
If you force American companies to spend 50-100x more than their competitors abroad here is what will happen:
1. American companies spending 50-100x will at some point become financially impaired and the stock market will crater. This is the equivalent of the US Government saying we can only buy oil at $800/barrel even when the free market sells the same oil for $80/barrel.
2. In the short term, the revenues of the Closed American Labs may remain hi. But then their revenue will also crater because their customers in the US will be impaired and their customers abroad will have already flipped to cheaper solutions.
This would be a terribly self-defeating form of intervention if it were to happen.
MTS: SITUATION BREWING: The Trump administration is considering restricting cutting-edge Chinese AI models, with momentum reviving after the launch of Kimi K3, per Axios.

Re @vorfluxai is self-serve. Many people didn't realize it. You can go sign up and try it.

If you want Gbrain, but don’t want to fiddle around with Hermes or OpenClaw, http://gbrain.io will have a brain up and running in a few minutes and each brain runs on a dedicated server.
Garry Tan: GBrain is a free open source retrieval library I give away that works best with Hermes Agent and OpenClaw but can work with Codex and Claude Code as well
It powers my own personal company brain and personal ai
https://github.com/garrytan/gbrain

Introducing Silent Speech interface for your existing devices!
From your phone or computer, it lets you communicate with AI at the speed of voice, using only your mouth movements, no sound.
Demo 0:13 and 1:09

Amjad Masad on Going Direct & Founder Storytelling
Replit CEO Amjad Masad joins a16z's Erik Torenberg on how he went from crippling stage fright to one of the most famous CEOs on X, which platforms actually matter and why, and what founders get wrong about building in public.
00:50 Memeing a dream into reality
03:02 Conquering stage fright
04:00 Make your mistakes early
05:23 Building in public
06:41 The authenticity edge: Elon, Zuck, Trump
08:42 Become uncancellable
10:47 Advice for Anthropic and OpenAI
15:09 The day AI deleted prod
17:39 Who should go direct?
19:37 X vs Instagram vs YouTube
22:52 Go direct role models
@amasad @eriktorenberg

Use one agent to do the work and another to review it against a rubric.
@trq212 explains how this works for video shorts:
“For something like, ‘Is this a good video short or not?’ you don’t have a deterministic answer.
You should have a separate verification agent read a rubric, review the short, and give feedback.
We call it self-preferential bias. When a model prefers its own output, it’s going to be more lenient at verifying it.”
📌 Watch the full episode here: https://youtu.be/aVO6E181cNU
Peter Yang: “We’ve got /loop, /goal, and workflows. They all get the agent to run for long periods of time.”
Here’s my new episode with @trq212 from the Claude Code team, where he shared how he plans, builds, and runs loops with Claude Code:
→ How /loop, /goal, and workflows work

Replit CEO Amjad Masad on why going direct wins:
"Whenever there's a piece of communication or a story I wanna tell, and there's a choice between public and private, I might bias towards public. I often write something publicly and then paste it on Slack."
"Being yourself is kind of a big radical thing that maybe started happening with Elon maybe 10 years ago. CEOs were supposed to be this very polished thing. They had to be filtered through this prism of PR, and journalists were kind of the way to get to the public."
"Zuck changed the entire narrative around him by just being himself and actually talking directly to people."
@amasad @eriktorenberg
a16z: Amjad Masad on Going Direct & Founder Storytelling
Replit CEO Amjad Masad joins a16z's Erik Torenberg on how he went from crippling stage fright to one of the most famous CEOs on X, which platforms actually matter and why, and what founders get wrong about building in public.

The most scalable thing you can do right now may look completely unscalable.
As @garrytan puts it: don’t scale the team yet. Scale the work.
Don’t do one-off work. Skillify it. 100x productivity gains are achievable if you build the right systems.
Must-watch: https://www.youtube.com/watch?v=eBUyTS7SzV4

Amjad Masad, CEO of Replit, on why founders need to go direct on every platform:
"X is very good for Silicon Valley, for tech, for in-group messaging, for maybe one hop outside of Silicon Valley, which is the early adopter. Once you want to go to late adopters or early majority, Twitter becomes actually not very impactful there."
"Instagram, Facebook, YouTube is where you get to the early majority. I just started focusing on it in maybe February or something like that, and we're getting an insane amount of views."
"I would start with Twitter because that's where the news starts. This is where you can have outsized influence."
@amasad @eriktorenberg
a16z: Amjad Masad on Going Direct & Founder Storytelling
Replit CEO Amjad Masad joins a16z's Erik Torenberg on how he went from crippling stage fright to one of the most famous CEOs on X, which platforms actually matter and why, and what founders get wrong about building in public.

Like good odds I'm wrong but I wanted to write this down:
It's pretty clear to me that superintelligence is here and it's more powerful than us and it's moving where things are going now, not humans anymore
I don't see many people realize this yet, it feels like that pic I posted the other day, everyone is running after the same carrot which is the AI, thinking they're special, and their work is special and their use of AI is special, but it's really not, we're all mostly making the same slop, I mean it's nice slop, useful slop but everyone is making the same slop
And because it's so fast to make things, like people used to spend a year on just building an app, now it's done in hours, it's such an immense change, people send me their projects and it all looks the same, it's useful but it's slop
Non-technical people (e.g. the gfs) are now building the same or better things than technical people like us
So AI has made everybody is just as capable as everyone else, it equalized everyone in the world, as in everyone can make everything (software, music, images, art etc) now and everyone is equal, at least in the digital realm now
And maybe now we're in some odd transitionary time where we have to find out waht the next differentiator is as coding/building/execution isn't one anymore for sure
I thought it'd be distribution but who knows, obviously creativity and ideas, but if you can copy a successful apps in an hour, then how does that differentiating work?
A guy on here @aporia9n wrote how he increasingly meets founders who blast through apps/products/startups, kind of money grabs, very quickly jumping on a trend, building super quickly with AI, make lots of money quick, everyone copies them, then their margins go to 0%, and they shut it down and go to the next thing, in a way they found one differentiator which is speed
That's one way to do it, but it shows how radically things are changing I think
For now I think the only ones winning is the AI superintelligence itself and the companies providing the AI
@levelsio: I increasingly have no idea about anything anymore, time is changing too fast, and with AI everyone is equal now, nobody really knows anything while there is a superintelligence that's actually moving things into a direction that nobody can predict

Jessica modeling the rarest YC shirt of all. I wondered why I'd never seen these in the wild. It's because she had them made just for her in various colors in 2006.

Great summary from @garryslist about corruption in District 10 — SFPUC “Friends with Community Benefits” scheme, YCD, Dwayne Jones, mentor to Shamann Walton, indicted on 59 counts of fraud, to Walton’s mentee and endorsed successor DJ Brookter. https://garryslist.org/posts/the-nonprofit-behind-san-francisco-s-pay-to-play-scandal-wants-another-supervisor-seat

Banning open-source AI would hurt defenders 10x more than attackers, which would make the world 10x more dangerous and this is a good example why!

No tax on tips. No tax on overtime. No tax on Social Security. Except: all taxed with narrow carve-outs, and all of it expires end of 2028.⁣
⁣
The one part of the tax bill that doesn't expire? Tax cuts for the wealthy. Those are permanent.⁣
⁣
Populism with an asterisk.⁣
⁣
https://newsletter.platypuseconomics.com/p/justin-answers-your-questions-on

Amjad Masad on how to build a founder brand that keeps you alive
a16z: Amjad Masad on Going Direct & Founder Storytelling
Replit CEO Amjad Masad joins a16z's Erik Torenberg on how he went from crippling stage fright to one of the most famous CEOs on X, which platforms actually matter and why, and what founders get wrong about building in public.

Things upper middle class parents teach their children that I've learned as an adult:
- Talking your book is not tacky by default
- There's an entire game to getting into your institution of choice. Learning & playing that game is ok and even encouraged
- Admitting every error you make is unnecessary
- There are specific career paths to follow if you want to make a lot of money (tho now less true in age of AI)
- "Networking" is more like making friends
- Relationships are how things actually get done
- Ask for the things you want
- Rules are defaults and often aren't even written for you specifically
- Really the biggest that encapsulates them all: all systems are all made of people, with their own incentives, desires, and goals

LA Unified just signed union contracts so expensive the county says its main fund will hit a negative $1.46 billion balance by FY 2027-28.
The nation’s second-largest school district is on track to run out of money because it prioritized union contracts over students and staff.

🦔A Nikkei investigation found that Alphabet, Microsoft, Amazon, Meta, and Oracle have $1.65 trillion in debt that doesn't appear on their balance sheets, more than the $1.35 trillion they officially report. These are GPU contracts, data center leases, and joint ventures that don't count as debt under accounting rules until the facilities go live. Meta's hidden debt is $420 billion, triple its reported debt. Oracle's grew 30-fold in four years. All five declined to comment.
My Take
Nikkei examined the actual filings and put a number on something the BIS already flagged as "shadow borrowing" back in March. These companies owe more off their balance sheets than on them, and the accounting rules let them keep it that way until the data centers go live. That's legal, but it means investors looking at quarterly earnings this week are seeing less than half the picture.
Four of these five report earnings in the next two weeks. The reported debt will look manageable. The $1.65 trillion in footnotes won't make the headlines. But when those data centers start operating, the leases hit the books all at once. If AI demand comes in below projections, those facilities get marked down and the losses land on the investors and insurance policyholders who funded the construction through private credit and project bonds without realizing how much total exposure they were carrying.
Hedgie🤗