IronTri99

The frequency of avalanches is inversely proportional to their size.

[IronTri99 @IronTri99]

$PLTR fits right in. Bullish.

[vas @vasuman]

A former COO of Citadel Securities is telling you where the money is in AI

For free

[IronTri99 @IronTri99]

@elonmusk Open AI. Zuck. They’re theives with zero original ideas of their own.

[IronTri99 @IronTri99]

Oops

[*Walter Bloomberg @DeItaone]

IRAN HAS TOLD U.S. OFFICIALS THAT RECENT ATTACKS ON SHIPPING WERE FROM 'AN ERRANT PART OF THEIR SYSTEM' – SENIOR U.S. OFFICIALS

[ₕₐₘₚₜₒₙ @hamptonism]

This 57 minute conversation with the founder of Palantir is literally worth more than reading 20 books:

[IronTri99 @IronTri99]

$15/sh added to the stock price isn’t a nervous breakdown. Idiots.

[Keith Fitz-Gerald @fitz_keith]

Here's my take on Karpus Maximus on CNBC...

Critics are painting his appearance as a nervous breakdown or worse.

Hardly.

First, Karp clearly laid out what he sees ahead, and the inherent fallacy associated with frontier AI models.

Karp was driving at something I run into all the time… the world’s best and most successful executives know that protecting their business model is critically important.

Competitors – meaning every other AI model out there that Karp harped on – require that you load UP your data, design prompts that feed into somebody else’s models, which means they get to see how your business operates and what you deem as important, before feeding that back to you with specific results.

Not tracking?

This would be like Coca-Cola being required to hand its secret formula to a lab that also formulates recipes for Pepsi — and then being forced to trust that lab to keep quiet about what it sees while paying for the privilege.

Palantir, on the other hand, keeps everything private and is entirely auditable every step of the way.

Second, nearly everything he said went in one ear and out the other around the table.

I found the blank looks priceless.

The average critic cannot be bothered to understand the difference between what Palantir does and what “competitors” do so they wind up parroting false information, misleading narratives and half-baked thinking.

Third, Karp confirmed what I’ve been saying all along with hard, unambiguous numbers.

I’m paraphrasing but here ya go… “We have much more demand than we can supply. If you just look at our financials you can see, 2 years out, you can see $15-$18B in free cash flow.”

Think about that.

Karp said that Palantir could 4-5X in the next few years.

Exactly what I’ve said all along.

Perhaps I am naïve but I found it astonishing that not one person at the table had an apparent interest in exploring that line of thinking on anything other than a cursory level despite the fact that it is ostensibly their job. 🤦‍

I hope I own enough shares.

Keith’s Investing Tip: Founders know their companies better than anybody else so when they make statements that are grounded in fact and knowledge, smart investors would be wise to pay attention.

Trade idea: Obviously, own $PLTR as an investment. BUT get ready. Shares are now up ~23% off recent lows and my guess is that it’s only a matter of time before there’s another short attack. Tactically speaking, that presents a potential short-term play using putskies while premiums are cheap and volatility is comparatively lower than it’s been in a while.

If I’m correct, that’d be a GIFT to long term investors.

[David Sacks @DavidSacks]

Legacy Media types are calling this Alex Karp interview a “crash-out” so that’s your first clue that he is actually saying something extremely insightful. He is articulating what real “AI safety” looks like in the enterprise.

Not abstract alignment research or certification by a government-run DMV for AI. Real AI safety for businesses is the ability to control their own data, model weights, and compute — so a frontier lab can’t hoover up their proprietary knowledge and turn it into their next product.

As Karp explains, technical customers want “control over their compute, their models, their data stack, and their alpha. They want to know they own the means of production, and it’s not being transferred to someone else.”

Don’t think that can happen? Just look at Figma. According to The Information, Anthropic “blindsided” its then-business partner with the launch of Claude Design. Figma’s founder said Anthropic had not been “consistently honest” with them. Anthropic’s chief product officer had even served on Figma’s board until three days before the launch of Claude Design. Figma’s stock has fallen sharply this year while Anthropic’s valuation has surged.

This isn’t an isolated example. Anthropic has launched Claude Science, Claude Security, Claude Legal, and of course Claude Code — each expanding into categories previously served by companies building on top of their models. The pattern is consistent: watch where value is being created, then move in directly. Dominate the model layer, then use that position to capture the most lucrative verticals.

Dario has argued that open source models powerful enough to compete with Anthropic are “dangerous.” But dangerous to whom? Not to enterprises that want to retain control over their data and workflows. Dangerous to a business model that benefits from customers having few real alternatives at the model layer.

As Karp exposes, true enterprise safety isn’t trusting that a lab’s future roadmap won’t include your business. It’s retaining the ability to choose — at the model layer — who gets to see and use your alpha.

[Bill Ackman @BillAckman]

An important read on the hyperscalers and why we are bullish.

[Wynne @wynneflh]

Fished for years, finally learned why hooks always twisted 🤣☺️🤌

[Tom Nash @iamtomnash]

$PLTR has been getting punched in the face lately because the market has convinced itself that OpenAI and Anthropic are coming for Palantir.

I think that is moronic.

Saying LLMs are eating Palantir is like saying paint brush manufacturers are putting painters out of business…

OpenAI and Anthropic are building the brains. Palantir is building the nervous system that actually connects the brain to the body.

Big difference.

The model can tell you what it thinks. Palantir helps a company decide whether that answer is using the right data, whether it fits the right workflow, whether the right human signed off, and whether the action actually gets pushed into the real system where money, logistics, defense, fraud, supply chains, hospitals, or governments operate.

That is the part people keep missing.

LLMs are insanely valuable, but as the major models keep improving, they are also starting to look more similar. GPT, Claude, Gemini, Grok, all of them are getting better. The gap between them is no longer the whole story.

For enterprises, the real question is not just, “Which model gives me the smartest answer?”

The real question is:

Can I trust this answer?

Can I trace where it came from?

Can I control who approves it?

Can I plug it into my existing systems?

Can I use it without blowing up compliance, security, privacy, or operations?

That is where Palantir lives.

OpenAI and Anthropic are trying to build intelligence. Palantir is trying to turn intelligence into operational outcomes.

Those are not the same business.

One is the engine. The other is the factory floor, the dashboard, the permission layer, the audit trail, the operator, and the process that turns the engine into actual production.

So when people say, “What if OpenAI replaces Palantir?” my answer is simple:

That is like saying Ferrari replaces roads because Ferrari makes a great engine.

Good luck driving that thing through a hospital procurement system, a defense agency, or a Fortune 500 compliance department.

That is why I have been doubling down on Palantir recently.

Not because Palantir has no risk. It absolutely does. Valuation risk is real. Execution risk is real. Expectations are sky high.

But the competition fear is misunderstood.

The market is treating Palantir like it is competing with the model companies.

I think Palantir is the company enterprises call when they want to actually use the models without setting the building on fire.

[darkzodchi @zodchiii]

A Stanford team just published the 16-page PDF on “How to structure an AI agent”

Structure matters more than how you prompt it, and it's backed by hard numbers.

Build → Reflect → Curate → Reuse

• Build: the agent starts with a structured context, not a clever one-off prompt.

• Reflect: it watches what actually worked during execution, no labels needed.

• Curate: it folds those wins into an evolving playbook instead of a static prompt.

• Reuse: the next run starts from that refined structure, getting stronger each time.

This is exactly why senior engineers build the structure first in Claude Code, then let the agent run.

Read the paper, then grab the setup below 👇

[Avid @Av1dlive]

Google DeepMind just lost a Nobel Prize winner, a Gemini co-lead, and two AlphaFold researchers to Anthropic this week

$270B wiped from Alphabet.

in 42 minutes, Nenad Tomašev, their Senior Staff Research Scientist, laid out exactly what they disagreed on:

"we are not building AGI.
we are building a society of agents.

hundreds of thousands. millions of artificial decision makers."

i wrote the full guide on running a thousand of them while you sleep.

bookmark it before the economy forms without you.

[Codez @0xCodez]

A senior Anthropic engineer just dropped 11-page PDF on "Loop Engineering" for agentic systems.

The shift: you stop prompting the agent. You build the system that prompts it instead.

Schedule → Discover → Build → Verify → Repeat

Every loop runs one turn, five moves:

• Discovery: it finds its own work - failing CI, open issues, recent commits - instead of being handed a list.

• Handoff: each task gets an isolated git worktree so parallel agents don't collide.

• Verification: a second agent, told to assume the code is broken, reviews the first. The "thing that can say no."

• Persistence: results get written to disk, never left in a context window that gets flushed.

• Scheduling: an automation wakes it on a timer. That's what makes it a loop.

The key insight: an agent grading its own work always praises it.

This 11-page PDF changed how I'm building agentic systems today.

Read it now, then explore the article below.

[Paweł Huryn @PawelHuryn]

Claude Code creator, Boris Cherny:

"I don't prompt Claude anymore. I have loops prompting Claude and figuring what to do"

Instead of watching Netflix, check these 16 minutes of insights on building loops that prompt itself and the future of work.

Watch it, then read the full guide on loops below.

[Movez @0xMovez]

A senior Google engineer just dropped a 19-page PDF on "Loop Engineering" for LLM and agentic systems.

Act → Observe → Learn → Repeat

• Act: the LLM proposes a code transformation (tile this loop, parallelize that one).

• Observe: a compiler runs it and reports back - is it valid? faster? slower? by how much?

• Learn: the LLM reads that feedback and adjusts its next move.

• Repeat until it stops finding improvements.

The agent gets smarter purely from grounded feedback inside its own context window.

This 19-page PDF totally changed the way I’m building agentic systems today.

Read it now, then explore the article below.

[Ihtesham Ali @ihtesham2005]

Demis Hassabis confirmed every frontier AI lab is working on recursive self-improvement and in the same sentence said the safety risk of removing humans from the loop entirely keeps him up at night.

That combination should stop you.

The CEO of Google DeepMind just confirmed that the thing most people treat as a theoretical future risk is already the active focus of every serious lab on earth right now.

He explained why it works in coding and math. The feedback loop is fast. You can verify whether an answer is correct almost instantly. You can generate synthetic training data from it. The loop closes quickly and cleanly.

Then he said where it breaks down.

In biology, chemistry and physics. Any domain where verifying a hypothesis requires a physical experiment in the real world. The loop does not close in seconds. It closes in weeks or months.

Geoffrey Hinton said in his Nobel lecture that recursive self-improvement is the development he fears most and that once started it may not be possible to stop. Hassabis is not pushing back on that. He is describing the guardrails labs are building around a process they are already running.

Every lab has to think carefully about the safety of a process where no human is in the loop.

He said that as a constraint they are navigating right now.

The question they are sitting with is how much of it to let run without a human watching.

(Watch the full interview on YouTube at @twominutepapers channel)

[Anatoli Kopadze @AnatoliKopadze]

Anthropic engineers just showed how they build a full app from scratch, using a loop of agents

40 minutes from the team behind Claude Code

they used three agents: one to plan, one to build, one to judge, cycling until the app actually works

the winners won't have the smartest model, they'll have the best loop

watch it, then read the full guide on how to actually use loops below

[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 app in a day instead of 2 weeks

in the right hands, this changes everything:

[IronTri99 @IronTri99]

@elonmusk It’s like when the SETI@home project used idle web connected computers to process radio signals in the search for signs of extraterrestrial intelligence. This is no different except it’s super fast.

[Antonio Lupetti @antoniolupetti]

"Transformers" by Daniel Jurafsky and James H. Martin is one of the clearest and most mathematically grounded introductions to the Transformer architecture I have ever read.

Chapter 8 introduces the Transformer as the standard architecture behind modern large language models. What makes this chapter particularly interesting is its step-by-step presentation of the underlying mechanisms: contextual embeddings, self-attention, query, key and value vectors, scaled dot-product attention, multi-head attention, residual streams, feedforward layers, layer normalization, masking, and the parallel matrix formulation of attention.

In particular, the treatment of attention as a weighted sum of contextual representations is especially valuable. The chapter first develops an intuitive, simplified view of attention and then gradually derives the full formulation using the Q, K, and V matrices. This approach makes it easier to understand what is actually happening inside the architecture from an algebraic and matrix-based perspective, rather than simply viewing the usual block diagrams.

I think it is an excellent resource for anyone interested in understanding how Transformers work from linguistic, mathematical, and computational perspectives.

https://t.co/3fitdPy6Fv

[Iqra @AiWithIqra]

THAT'S WHY AIRLINES HATE CLAUDE 4.6

Flight for $879. I paid $299.
No points. No affiliations. No VPN.

Here are 8 prompts I used to travel like a pro↓