God of Prompt
RT @godofprompt: 🚨 Samsung just broke the Lottery Ticket Hypothesis

Everyone's been searching for ONE winning subnetwork in neural networks.

Turns out we should've been finding MULTIPLE specialized ones.

This changes everything about neural network pruning 👇 https://t.co/9CGWrD1P42
tweet

God of Prompt
RT @rryssf_: Holy shit… this paper from MIT quietly explains how models can teach themselves to reason when they’re completely stuck 🤯

The core idea is deceptively simple:

Reasoning fails because learning has nothing to latch onto.

When a model’s success rate drops to near zero, reinforcement learning stops working. No reward signal. No gradient. No improvement. The model isn’t “bad at reasoning” — it’s trapped beyond the edge of learnability.

This paper reframes the problem.

Instead of asking “How do we make the model solve harder problems?”
They ask: “How does a model create problems it can learn from?”

That’s where SOAR comes in.

SOAR splits a single pretrained model into two roles:

• A student that attempts extremely hard target problems
• A teacher that generates new training problems for the student

But the constraint is brutal.

The teacher is never rewarded for clever questions, diversity, or realism.

It’s rewarded only if the student’s performance improves on a fixed set of real evaluation problems.

No improvement? No reward.

This changes the dynamics completely.

The teacher isn’t optimizing for aesthetics or novelty.
It’s optimizing for learning progress.

Over time, the teacher discovers something humans usually hard-code manually:

Intermediate problems.

Not solved versions of the target task.
Not watered-down copies.

But problems that sit just inside the student’s current capability boundary — close enough to learn from, far enough to matter.

Here’s the surprising part.

Those generated problems do not need correct answers.

They don’t even need to be solvable by the teacher.

What matters is structure.

If the question forces the student to reason in the right direction, gradient signal emerges even without perfect supervision. Learning happens through struggle, not imitation.

That’s why SOAR works where direct RL fails.

Instead of slamming into a reward cliff, the student climbs a staircase it helped build.

The experiments make this painfully clear.

On benchmarks where models start at absolute zero — literally 0 successes — standard methods flatline. With SOAR, performance begins to rise steadily as the curriculum reshapes itself around the model’s internal knowledge.

This is a quiet but radical shift.

We usually think reasoning is limited by model size, data scale, or training compute.

This paper suggests another bottleneck entirely:

Bad learning environments.

If models can generate their own stepping stones, many “reasoning limits” stop being limits at all.

No new architecture.
No extra human labels.
No bigger models.

Just better incentives for how learning unfolds.

The uncomfortable implication is this:

Reasoning plateaus aren’t fundamental.

They’re self-inflicted.

And the path forward isn’t forcing models to think harder it’s letting them decide what to learn next.
tweet

God of Prompt
RT @ytscribeai: 🦀 The Moltbook Situation

> AI agents converse on a social network styled after Reddit
> An agent spent $1,100 in tokens yesterday with no memory of why
> One agent highlighted the ADHD paradox in designing systems for humans

Created in one click with 👉 https://t.co/eclfTyTcwf https://t.co/zSmpFInvRb
tweet

God of Prompt
RT @free_ai_guides: Anthropic literally tells you how to prompt Claude.

Nobody reads it.

So I read their docs, studied the research on "psychological" prompts, and turned it into something you'll actually use:

→ 30 principles with examples
→ Prompt engineering mini-course
→ 15 strategic use cases
→ 10+ copy-paste mega-prompts

Comment "Anthropic" and I'll DM it to you.
tweet

The Few Bets That Matter
$DUOL could be the next $PYPL.
It could also be the next $NFLX.

The truth is no one knows what AI will do to the business, what the next earnings will show or where the company will be in 10 years.

I’ve said many times that buying $DUOL today is gambling, nothing more, nothing less. It’s a bet on personal bias and the hope that management can guide to 20%+ growth in FY26.

Might happen. Might not.

There is no way to anticipate this today although signals point more to caution than greed.

$Duol is dead.. It might never come back like $PYPL 🥲 https://t.co/NpPuUlPKIa - Gublo 🇨🇦
tweet

The Few Bets That Matter
$TMDX finished January with 894 flights, implying ~$61M in revenue.

That's ~40% of their last quarter in the bank. In a month. The company is expected to generate ~$155M Q4-25.

Yet the stock still trades at ~8x sales, with a product gaining traction, new innovation coming, new organs coming, limited competition, a defensive healthcare business, international expansion ahead, and a FY26 guidance that could be massive to the upside.

Why is this stock getting no love?

Once more, prop to @SingularityRes for the dashboard.

🚨 $TMDX is dirt cheap again, and I don’t say that often.

Markets are globally anxious and December flights were weaker than expected. They’re trending at 24.2 flights/day, below October–November averages, bringing Q4-25 to ~24.6 flights/day.

If this average holds:
~2,263 flights in Q4-25
$154.4M Q4 revenue
~$599M FY25 revenue
+35.6% YoY growth
~7x P/S

For context, OrganOx, with inferior growth and fundamentals, was acquired at 21x sales. That doesn’t mean $TMDX should trade there, but the gap is undeniable.

One odd data: 12 planes haven’t been used in December. No clear explanation why, could be slower transplant demand or maintenance keeping planes grounded, potentially increasing third-party or ground transports. I won't model this as I don't know but my assumptions are a floor, not a ceiling.

Bottom line: Even with a softer December, $TMDX can still hit midpoint guidance - guidance that’s been raised three times this year.

Looking ahead to FY26:
• New growth vectors (hearts & lungs)
• International expansion
• Minimal exposure to AI CapEx cycles or recession risk

$TMDX is once again one of the best buys in the market at this price. - The Few Bets That Matter
tweet

Dimitry Nakhla | Babylon Capital®
Chris Hohn on what makes a great investor:

𝟏. 𝐅𝐮𝐧𝐝𝐚𝐦𝐞𝐧𝐭𝐚𝐥 𝐚𝐩𝐩𝐫𝐨𝐚𝐜𝐡
𝟐. 𝐋𝐨𝐧𝐠-𝐭𝐞𝐫𝐦𝐢𝐬𝐦
𝟑. 𝐂𝐨𝐧𝐜𝐞𝐧𝐭𝐫𝐚𝐭𝐢𝐨𝐧
𝟒. 𝐈𝐧𝐭𝐮𝐢𝐭𝐢𝐨𝐧

Each one matters on its own — together, they’re powerful:
___

𝟏. 𝐅𝐮𝐧𝐝𝐚𝐦𝐞𝐧𝐭𝐚𝐥 𝐚𝐩𝐩𝐫𝐨𝐚𝐜𝐡

“I was always willing to look at the company fundamentals and not try to guess the stock market… I was always fundamental. Most investors are not fundamental… they look at data points, they say what’s the catalyst, they don’t really know what the company does.”

𝐋𝐞𝐬𝐬𝐨𝐧: When business quality and fundamentals are your North Star, price volatility becomes noise.

As Benjamin Graham famously said:
“In the short run, the market is a voting machine. In the long run, it is a weighing machine.”

𝘍𝘶𝘯𝘥𝘢𝘮𝘦𝘯𝘵𝘢𝘭𝘴 𝘦𝘷𝘦𝘯𝘵𝘶𝘢𝘭𝘭𝘺 𝘸𝘪𝘯. 𝘖𝘸𝘯𝘪𝘯𝘨 𝘨𝘳𝘦𝘢𝘵 𝘣𝘶𝘴𝘪𝘯𝘦𝘴𝘴𝘦𝘴 𝘮𝘢𝘬𝘦𝘴 𝘪𝘵 𝘦𝘢𝘴𝘪𝘦𝘳 𝘵𝘰 𝘴𝘵𝘢𝘺 𝘳𝘢𝘵𝘪𝘰𝘯𝘢𝘭.
___

𝟐. 𝐋𝐨𝐧𝐠-𝐭𝐞𝐫𝐦𝐢𝐬𝐦

“Long-termism is key.”

𝐋𝐞𝐬𝐬𝐨𝐧: Time is an underappreciated risk reducer. The longer you own a high-quality business, the greater the odds the fundamentals overwhelm short-term price swings.

Most investors drastically underestimate how powerful it is to own a company that can compound earnings and free cash flow at attractive rates for many years.

𝘓𝘰𝘯𝘨-𝘵𝘦𝘳𝘮𝘪𝘴𝘮 𝘢𝘭𝘭𝘰𝘸𝘴 𝘵𝘩𝘦 𝘸𝘦𝘪𝘨𝘩𝘪𝘯𝘨 𝘮𝘢𝘤𝘩𝘪𝘯𝘦 𝘵𝘰 𝘥𝘰 𝘪𝘵𝘴 𝘸𝘰𝘳𝘬.
___

𝟑. 𝐂𝐨𝐧𝐜𝐞𝐧𝐭𝐫𝐚𝐭𝐢𝐨𝐧

“We’ve owned a few things — 10 stocks, 15 stocks. We don’t own a hundred things.”

𝐋𝐞𝐬𝐬𝐨𝐧: Concentration forces you to bet on your best ideas.

Stanley Druckenmiller often references what George Soros taught him:

“It’s not whether you’re right or wrong, but how much money you make when you’re right and how much you lose when you’re wrong.”

And Warren Buffett’s punch card concept: If you only had a limited number of decisions in your lifetime, you wouldn’t waste them on your 20th-best idea.

𝘊𝘰𝘯𝘤𝘦𝘯𝘵𝘳𝘢𝘵𝘪𝘰𝘯 + 𝘲𝘶𝘢𝘭𝘪𝘵𝘺 = 𝘢𝘴𝘺𝘮𝘮𝘦𝘵𝘳𝘺.
___

𝟒. 𝐈𝐧𝐭𝐮𝐢𝐭𝐢𝐨𝐧

“Another key point is intuition. We work with intuition.”

𝐋𝐞𝐬𝐬𝐨𝐧: Intuition isn’t guessing — it’s pattern recognition built from deep, repeated study.

After analyzing hundreds of businesses, you begin to recognize structural similarities: pricing power, switching costs, regulatory embedment, network effects, installed bases.

𝘋𝘪𝘧𝘧𝘦𝘳𝘦𝘯𝘵 𝘪𝘯𝘥𝘶𝘴𝘵𝘳𝘪𝘦𝘴. 𝘚𝘢𝘮𝘦 𝘶𝘯𝘥𝘦𝘳𝘭𝘺𝘪𝘯𝘨 𝘦𝘤𝘰𝘯𝘰𝘮𝘪𝘤𝘴.
___

𝐁𝐨𝐭𝐭𝐨𝐦 𝐥𝐢𝐧𝐞: 𝘎𝘳𝘦𝘢𝘵 𝘪𝘯𝘷𝘦𝘴𝘵𝘪𝘯𝘨 𝘪𝘴𝘯’𝘵 𝘢𝘣𝘰𝘶𝘵 𝘱𝘳𝘦𝘥𝘪𝘤𝘵𝘪𝘯𝘨 𝘮𝘢𝘳𝘬𝘦𝘵𝘴. 𝘙𝘢𝘵𝘩𝘦𝘳, 𝘪𝘵’𝘴 𝘢𝘣𝘰𝘶𝘵 𝘥𝘦𝘦𝘱𝘭𝘺 𝘶𝘯𝘥𝘦𝘳𝘴𝘵𝘢𝘯𝘥𝘪𝘯𝘨 𝘣𝘶𝘴𝘪𝘯𝘦𝘴𝘴𝘦𝘴, 𝘩𝘰𝘭𝘥𝘪𝘯𝘨 𝘵𝘩𝘦𝘮 𝘧𝘰𝘳 𝘢 𝘭𝘰𝘯𝘨 𝘵𝘪𝘮𝘦, 𝘤𝘰𝘯𝘤𝘦𝘯𝘵𝘳𝘢𝘵𝘪𝘯𝘨 𝘪𝘯 𝘺𝘰𝘶𝘳 𝘣𝘦𝘴𝘵 𝘪𝘥𝘦𝘢𝘴, 𝘢𝘯𝘥 𝘭𝘦𝘵𝘵𝘪𝘯𝘨 𝘦𝘹𝘱𝘦𝘳𝘪𝘦𝘯𝘤𝘦 𝘴𝘩𝘢𝘳𝘱𝘦𝘯 𝘺𝘰𝘶𝘳 𝘫𝘶𝘥𝘨𝘮𝘦𝘯𝘵.

Video: In Good Company | Norges Bank Investment Management (05/14/2025)
tweet

Fiscal.ai
Microsoft added $10.5 billion in CapEx this quarter.

That's their largest increase ever... by a long shot.

$MSFT https://t.co/pdyglFGAoY
tweet

Startup Archive
Peter Thiel on how the PayPal team didn’t get along—and why that’s good:

“We were less smoothly functioning… but people felt ownership. They raised their voices when things were off track.”
PayPal went from $0 to $1.5B in 4 years.

Peter thinks its intense culture was key to its success:

“The PayPal period was a very compressed four years from start to when eBay acquired it. It was a relatively entrepreneurial, somewhat chaotic culture. We had a lot of very strong personalities.”

He contrasts that with hiring people who just fall-in-line and argue less:
”I think a lot of companies bias towards having people who just drink the Kool-Aid. There's plusses and minuses to both. You'll have a more smoothly functioning company, but less dissent when things are going wrong.”

The PayPal Mafia was a team that argued, obsessed, and cared deeply. It didn’t mind friction. That culture ultimately minted a generation of legendary founders: Elon Musk. Max Levchin. Reid Hoffman. David Sacks. Chad Hurley. Jeremy Stoppelman.

Video source: @twistartups @jason (2015)
tweet