Shonenpepe 🫡

so it goes

[Taste Labs @taste_ai_]

tasteful team. join us

[ビーバーズ・ハイブ か8@NT金沢 @beavers_hive]

１００万回いいねされた猫、3Dモデル化🐱
3Dプリンターで印刷してみようかしら

[OneShotFox🇳🇿 @OneShotFox]

Generational mog during national anthem

[The Cats 𝕏 @TheCatsX]

A girl was streaming and her cat started reading the chat

[Will Manidis @WillManidis]

“Always go to the funeral”

[Midjourney @midjourney]

Announcing a new division of Midjourney called "Midjourney Medical"

[9GAG ❤️ Memeland @9GAG]

the monster that’s under your bed

[tobi lutke @tobi]

This is very good

[prayingforexits 🏴‍☠️ @mrexits]

Feels like so much of the hate around technology boils down to this one Dostoyevsky quote

[Will Manidis @WillManidis]

Dario (48 hours ago): “US gov should be able to block model deployment”

USG: *export controls models*

Dario: “not like that”

[Prompter @PromptLLM]

Insane take from Fable 5

[Eli Mernit @mernit]

One piece of advice we got during YC was to explain our company using verbs instead of nouns.

Early on, I walked into a meeting and did the opposite:

“We’re building a cloud platform for AI”

No one knew that that meant, their eyes glazed over. Then I started saying this instead:

“We containerize your code and run it on GPUs in the cloud so you don’t have to manage the infra yourself”

That clicked way more. Our brains understand verbs because they’re more concrete. If you describe your company using nouns, you risk people not understanding you.

And no one buys or invests in things they don’t understand.

[ol’ stocky ⛳️ @oldstocky]

I can understand why Italians saw this and decided to invent unemployment

[klöss @kloss_xyz]

I prompted Claude Fable 5 to use Python to generate a 9:16 social video and render it using ffmpeg. Told it to put its own personal spin on it so it’s aligned with Anthropic’s launch and to fully express what it’s like to be an LLM hated by Theo from its POV.

It made this lmfao.

[ℏεsam @Hesamation]

3Blue1Brown’s new video explains why every LLM is actually a compression machine.

everyone describes pre-training as “next token prediction” but that’s just the surface-level objective.

in reality it is a means to making the most efficient text compressor.

prediction and compression are two sides of the same coin.

when you train the model to predict the next token you’re not just teaching it to guess the next word but how to best encode the human knowledge it sees.

better compression
means better abstraction
means better reasoning

at some point, compression stops looking like storage or a database (as some like to call it on X)
and looks like an approximation of understanding.

[D2 Finance @D2_Finance]

This is not a robotics thesis. It is a funnel.

Slick video, charismatic marketer, “why I bet my career on robotics.” He works for $BOT. RoboStrategy, the closed-end vehicle Andrew Kang now runs and openly brands the “MicroStrategy of robotics.”

Same playbook: a committed equity facility that only works if the share price stays bid and retail keeps showing up to absorb the issuance. The content is the top of the funnel. You are the liquidity.

We traded millions of dollars of converts on the real robotics names in the late 2010s, mostly within a convertible arbitrage mandate, not fundamental stock-picking. Which is exactly why we can point at the honest alternatives here, the ones nobody is cutting a hype video for, because there is no salesman making big bucks dumping them on you.

If you want robotics exposure, the honest version is already listed in Tokyo. FANUC (6954), 20%+ operating margins through full vertical integration. Keyence (6861), asset-light, ROE most software firms would kill for. Yaskawa (6506), the servo muscle behind half the world’s arms. The Japanese Big Five ship 40%+ of global industrial robots. These are priced for competence, audited for decades, and you can size them without praying a NAV premium holds. Dull. Real. Yours at a clearing price.

And if you actually want the convexity, the asymmetric leg, it is one layer down in the actuation supply chain, not in a closed-end fund. Harmonic Drive Systems (6324), Nabtesco (6268), the strain-wave reducer makers. A humanoid needs roughly 20 harmonic drives per unit, which makes precision reducer supply the binding constraint on the entire scaling story. That is where the torque is. That is also where you get your face removed if the shipment curve disappoints. Omdia has 2025 humanoid shipments up ~480% to ~13,000 units, real growth, and a rounding error against the valuations now leaning on it.

So here is the actual choice. You can own the cash-generating incumbents directly, at a price, with no wrapper skimming you. You can own the actuation convexity directly, eyes open, and underwrite the humanoid curve yourself. Or you can buy a single ticker at a premium to its own book, structured so the manager can print on your ‘enthusiasm’, and call that a career bet.

Robotics is real. The compounding is real. The funnel is also real, and it is looking for your money.

The edge was never finding the names. The edge is knowing who is selling, and to whom.

[Charly Wargnier @DataChaz]

🚨 ANOTHER MASTERCLASS FROM @3BLUE1BROWN

The compressibility of language isn’t just a math curiosity, it’s the hidden engine behind every LLM you use.

Grant’s new video reframes Shannon’s entropy through one elegant lens:

Prediction IS compression.

→ The better you predict the next word, the fewer bits you need to store it
→ Shannon measured English at ~1 bit per character: astonishingly compressible
→ This is exactly what GPT-style models optimize
→ Intelligence, in this framing, is compression

FUN FACT: Von Neumann told Shannon to name it “entropy” because nobody truly understands it anyway 😄

Decades later, that same concept became the bedrock of modern AI.

Deep-dive resources in the 🧵 ↓

[Laura Elliott @TinyWriterLaura]

i love things like this. people have always been people