F

[F @francogoat]

@aubreystrobel @tsmidnights13 they send them to rehab

[luca° @lucacs]

easily the best research account you’re somehow not following yet

[Bull Theory @BullTheoryio]

🚨JAPANESE AI STARTUP JUST MATCHED CLAUDE FABLE 5 AND MYTHOS PERFORMANCE.

Japanese AI lab just launched Fugu, a model trained to command other models.

Sakana AI, a Tokyo-based AI startup, was co-founded by researchers including one of the authors of the original Transformer paper, the blueprint that basically every AI model today is built on.

Sakana Fugu is a multi-agent AI system that feels like a single model, automatically orchestrating specialized models behind one API endpoint.

Fugu Ultra is claimed to match Anthropic's Fable 5 & Mythos Preview on the hardest engineering/science/reasoning benchmarks — and to beat Gemini 3.1 Pro, Opus 4.8 & GPT-5.5 on tasks like AutoResearch, mechanical design & financial forecasting.

Everyone's racing to scale ONE giant model but Fugu flips it: an LLM that orchestrates a pool of the world's best models (choosing who does what, delegating, verifying and merging it into one answer), even calling itself recursively.

Think of it as a conductor. You send one request to one API, and Fugu figures out the move: answer it solo, or assemble a squad of expert models and run the whole thing for you.

Sakana AI launched two models:
Fugu - fast, low-latency, everyday coding/chat
Fugu Ultra - max quality on hard, multi-step problems (AI research, paper reproduction, cybersecurity, patent search)

Export controls are hitting frontier models now, but because Fugu's agent pool is swappable, if a provider restricts access, it routes around it.

Both Fugu and Fugu Ultra can be accessed through a unified API, with subscription plans for everyday users and pay-as-you-go pricing designed for high-volume and enterprise workloads.

[Peter Steinberger 🦞 @steipete]

This is becoming my favorite way to read Twitter. https://t.co/eykEElxzu7

[Midjourney @midjourney]

Announcing a new division of Midjourney called "Midjourney Medical"

[Evanss6 @Evan_ss6]

Also super bullish on the 10 to 20 year timeframe.

1. AI will replace or augment most knowledge work.

2. AI spend as a proportion of knowledge-work comp is minuscule today, close enough to zero.

3. AI will revolutionize fields beyond software: robotics, biotech, materials science, physics, defense/military, etc. David’s argument is that coding revenue falls off → other revenue doesn’t replace it in time → air pocket. Given (2), even if coding softens, penetration into the broader knowledge-work base is still coming. AI spend should broaden over time.

4. All of this requires tremendous amounts of memory and compute for inference. Falling compute costs (his #2) are also demand-expanding: cheaper inference makes more workloads economic, a Jevons-style effect.

5. Even if you don’t believe any of this, winning the AI race is seen as existential by the US. The government is already turning from regulator into stakeholder: OpenAI is in active equity talks with Washington, and the Intel and IBM stakes are done. Expect more direct state support across the leading labs (Anthropic, DeepMind and peers), whether through equity, contracts, or otherwise.

A key crux of his argument is that 1) app revenue funds the buildout and 2) hyperscalers exercise normal ROI discipline.

1. understates it. The buildout isn’t underwritten by coding revenue specifically; it’s a bet on aggregate compute demand across every use case, made by hyperscalers and labs, not by the app-layer software firms whose code moats could erode. That moat erosion is an app-layer story. It doesn’t subtract from infra compute demand, which is broadening (see 2) and getting cheaper to serve (see 4). A SaaS-multiple shakeout isn’t an infra-capex collapse.

2. may not hold either. Spending here probably won’t be as disciplined as a strict microeconomic context would imply, given the strategic and geopolitical stakes. That may break, or at least delay, the “market realizes → spending stops” step.

For these reasons, I suspect the tailwind remains strong. The 2028 election matters a lot for the buildout, future regulation, taxation/UBI, etc.

So much can still go wrong: safety, high energy costs, rates and credit, employment disruption, liquidity, issuance. Whether we get a boom/bust over the next few years is hard to say. History suggests that’s often what happens, but I don’t think this argument is sufficient to conclude it.

Coding is only the first killer use case. AI spend is tiny relative to the knowledge work wage pool, cheaper inference may expand demand faster than costs fall, and compute is becoming strategic infrastructure.

None of these arguments eliminate the need for eventual returns. It may simply make the cycle longer, less disciplined, and more political.

All in all, I’m very bullish on human (and machine) ingenuity united toward a common goal.

[Milk Road AI @MilkRoadAI]

This is WILD!

Ray Kurzweil, the futurist who predicted the internet, smartphones, and AI says aging ends by 2032 (Save this)

Kurzweil, now 78 years old, told a live audience that humanity will reach longevity escape velocity by 2032 and he explained exactly what that means with mathematical precision.

Right now, for every year you live, you get back approximately five months of life expectancy from medical and scientific progress meaning you are losing roughly seven months of net life per calendar year.

Longevity escape velocity is the threshold where that ratio flips, for every year you live, you get back a full year or more from scientific progress, meaning your biological clock starts running backward.

Kurzweil's prediction is that threshold hits by 2032 and beyond that point, you do not simply stop dying of aging, you actively get younger every year.

The mechanism is AI-driven drug discovery at a scale that was physically impossible five years ago.

By 2030, Kurzweil argues, AI will be able to take a biological problem, generate millions of potential drug candidates, screen all of them, and run trials on simulated digital populations compressing decades of clinical research into weeks.

This is already happening.

David Sinclair's lab at Harvard used AI to virtually screen 8 billion molecules against aging targets and is now preparing human trials moving from $400,000 gene therapies toward a $100 pill that can reset biological age by 50 to 95% in four weeks.

Sinclair has already demonstrated the ability to reverse aging in mammals restoring sight in mice with optic nerve damage and reversing Alzheimer's symptoms in lab models.

Kurzweil's track record is what makes the 2032 claim impossible to dismiss.

He predicted the internet's global dominance in 1990, the defeat of a world chess champion by a computer in 1998, pocket-sized devices as primary communications tools in 1999, and AI passing professional exams in the mid-2020s, all before anyone else was saying it publicly.

If you are under 60 and in reasonable health, his message is stay alive, stay healthy, and get to 2032.

The tools on the other side of that date will be unlike anything medicine has ever produced.

[Damir Akaza @Damir_Akaza]

Jane Street, one of the richest and most secretive firms in the world, paid him between $330,000 and $600,000 a year, and in just a couple of months he built an AI system that runs TRILLIONS of operations per second

"we just hired a kid... and he turned out to be a supercomputer in a human body" is what they're whispering now at Jane Street

a math genius who pushed supercomputers to their limit. now Wall Street's quant traders are in shock

in this hour-long lecture he breaks down how to use his machine to process trillions of data points

bookmark it right now and watch it instead of reels to learn how to do the same ↓

[Electra @Electrarythm]

Brainwash yourself

[Jaxon @jaxoncoder]

Your iPhone has a hidden map of everywhere you've been.

Mine had 307 entries in 2 months.

The map is buried behind Face ID.
Apple shows you a summary.
Law enforcement gets the full database.

Here's how to find it, see the number, and shut it off tonight:

[Quinten | 048.eth @QuintenFrancois]

Only ~5% of SpaceX stock is floating right now

~95% $SPCX is still locked

Most don’t realize bearish pressure often comes later, when insiders finally get liquidity

Unlock schedule below ⬇️

[Johan N. @rk8215]

$AMPG very good write up my @babyfolio. He likes what he sees. So do I.

"We got a founder-led company, with over 10% stake in the business. Insiders are buying shares, and the fundamental trajectory is inflecting: revenue is growing, EPS losses are narrowing, and gross margins are expanding sharply (hitting 48% in Q1).

On top of that, the backlog is strong and active, highlighted by the ongoing commercial conversion of that $40M LOI."

[ChinoAleman @chinoalemano]

Yeah, the bullish part is the zero water confirmation.

But here's something people are sleeping on:

Jagan (@Jagtheaiguy) and Venkat (@topideafactory) both have X accounts.

Right now all the $DGXX alpha drops on LinkedIn.

But if enough of us follow them here?

They might start talking on X instead.

And then nobody has to dig through LinkedIn for the good stuff.

Go follow them.

Let's bring the alpha home. 🧥

[Leonard Rodman @RodmanAi]

THIS IS ABSOLUTELY WILD 🤯

Jack Dorsey's new AI tool, Goose, is 100% FREE.

You type:

"Build me a website like YouTube."

And Goose gets to work on its own:

→ Creates the entire project
→ Writes all the code
→ Installs dependencies
→ Fixes errors automatically
→ Keeps going until it's working

The crazy part?

• No monthly subscription
• Runs on your own device
• Your code stays private
• Completely open-source

Just a few years ago, building software meant hiring developers or learning to code.

Now you can start with nothing but an idea.

We're entering a world where ideas are becoming more valuable than technical skills.

[Khairallah AL-Awady @eng_khairallah1]

this is f*cking gold

How to build your first AI agent (Full guide)

if I had this a year ago, I would've shipped my first agent in a day instead of 2 weeks

in the right hands, this changes everything:

[Satya Nadella @satyanadella]

https://t.co/vLmiBKTtX3

[vivek @itsreallyvivek]

https://t.co/ZhOgyq7Vgn

[Paul Graham @paulg]

How to Earn a Billion Dollars: https://t.co/WeWBUkKym6

[Etherealize @Etherealize_io]

Erik Voorhees: “ETH is still the king, and I don’t see it being dethroned"

The founder of ShapeShift and Venice AI is asked if Ethereum was a “sustainable ecosystem.” He replies:

“I think [Ethereum] is more than sustainable. I think it is the clear winner of the smart contract innovation. It actually wasn’t the first mover in smart contracts, but it was the first one to achieve any sort of scale with smart contracts. What’s most important about Ethereum isn’t so much the first-mover advantage as much as it is the network effect it has had since it was released.”

Erik continues:

“I think both Bitcoin and Ethereum have achieved a network effect that is close to unassailable. People have gotten distracted with some of these other L1s, but if you look at metrics like where the developers are and where stablecoin volumes are, these are hard to fake metrics that are very important. They’ve always been predominantly on Ethereum. It’s not even close. I’m glad that other people tried to build L1s. The process of innovation and competition is really important. But ETH is still the king, and I don’t see it being dethroned. It has had various scaling challenges — the patchwork of L2s and the UX problems between them sucks. But I have a suspicion that Base is going to end up becoming the predominant L2 on top of the predominant L1 of ETH and that vertical is going to be very powerful and very strong. So yes, I’m always bullish on ETH in the same way I’m always bullish on Bitcoin.”

However, Erik warns that if Base loses its permissionlessness it “will flounder and deserves to die”:

“Base has designed things very well. It has gotten a lot of adoption and very quickly became the major L2 even though it was not the first mover. I think it’s gaining a network effect pretty quickly. It obviously has a very powerful corporate ally in Coinbase, and to the degree that Coinbase does not abuse that privilege, that’s a very good privilege. Abuse here means: if Coinbase tries to exert control over base such that it loses its permissionlessness, then it will flounder and deserves to die. But Coinbase has been a very good actor in this regard, and they deserve a lot of credit for demonstrating the principles of decentralization and permissionless innovation in several parts of what they do. Obviously the centralized exchange is not that, but it’s not trying to be either.”

Source: @CoinDesk (Dec 2025)

[Martin Shkreli @MartinShkreli]

pharma GOATs (includes their teammates):

Wong
Edelman
Holman
Aghazadeh
Rothbaum
Chen
Baker
Ramaswamy

you want to ask the question "does a drug work" or "will a drug work", ask the people who have made the most money answering said question. don't ask an influencer