Saturnin Pugnet's Avatar

Saturnin Pugnet

@satpugnet

Founding member @Worldcoin πŸ—ΊοΈ Prev @Caltech/@Imperialcollege - Wonder wildly, fail often, love recklessly, and leave the campfire brighter than you found it πŸ”₯ https://satpugnet.com

75
Followers
233
Following
64
Posts
18.10.2024
Joined
Posts Following

Latest posts by Saturnin Pugnet @satpugnet

We plan in nouns, but we live in verbs. Happiness is found in the furniture of our days, not the architecture of our lives.

09.03.2026 16:00 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

The headcount required to build serious software is approaching zero.

Soon, the biggest competitive advantage in tech will simply be having fewer people who are more AI-native.

06.03.2026 17:00 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

The strong litmus test of ethics is whether you would destroy your life's work if your morals demanded it.

It is particularly relevant in the era of AI. This is no longer hypothetical for a lot of people.

02.03.2026 17:03 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
Post image

Nothing like some friendly LLM feedback about your ideas to end the week

02.03.2026 02:35 πŸ‘ 0 πŸ” 1 πŸ’¬ 0 πŸ“Œ 0

The arrival of a superior entity is far less disruptive when you never claimed to be the pinnacle of the universe to begin with.

24.02.2026 10:32 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

The spiritual shockwave of AI will hit the West much harder than the East. Western culture centers reality around the individual; Eastern thought weaves humanity into the fabric of nature.

24.02.2026 10:32 πŸ‘ 0 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0

Too often people let early wins crystallize into their personalities. The early win becomes the story, the story becomes the identity, and the identity becomes the cage.

23.02.2026 17:00 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

I give it 2 years before a new martial art focused on sparring with robots emerges.

16.02.2026 00:35 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
Video thumbnail

Built a wealth OS for myself over the weekend.

Had the best time mansplaining FIRE to my gf on Valentine's day.

PS: Obviously all the data is mocked and randomized but I run the real deal on my laptop. With live connection to all my banks, brokerage accounts, prediction market accounts etc...

15.02.2026 19:39 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

Just realized today that the word for "Strike" in French is "Manifest"-Γ©.

Some jokes really just write themselves.

19.01.2026 16:59 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

Whether it is a dating app, a fusion reactor, or a spreadsheet, the principle follows the laws of physics and biology:

Value is created when the energy required to achieve an evolutionary goal is minimized.

14.01.2026 12:18 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

The successful startup drastically reduces friction to satisfy these drives. By minimizing resistance, the company increases the magnitude and velocity of these vectors.

14.01.2026 12:18 πŸ‘ 1 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0

It identifies a constant and fundamental evolutionary vector, either fear or one of the 7 Deadly Sins, for humans or for companies. We consider these similar because companies act like collective biological organisms.

14.01.2026 12:18 πŸ‘ 0 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0

In the long term, a successful company is simply an Efficiency Engine for Human Nature.

14.01.2026 12:18 πŸ‘ 0 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0

This morning, I was trying to find the single best mental model for what a company actually is. I wasn't looking for the MBA definition, I wanted the first-principles definition.

After thinking about the fundamental mechanics for a few hours, here is what I came up with:

14.01.2026 12:18 πŸ‘ 0 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0

Features my fingerprints to confirm that it is not AI.

14.12.2025 01:58 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
Post image

I think I might have made a mistake

14.12.2025 01:58 πŸ‘ 1 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0

We often talk about the chip war, but I believe the war for the training set is also happening right now, mostly invisible to the average user. These "glitches" are likely just the tip of the iceberg.

I wonder if anyone is actually tracking this shift in token distribution?

07.12.2025 14:05 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

By ensuring their data comprises a significantly larger percentage of the global training corpus, they can subtly influence how future "neutral" models process information.

07.12.2025 14:05 πŸ‘ 0 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0

As LLMs rely heavily on the open web for training, the composition of that data dictates the model's behavior. I believe China (or state-aligned actors) may be strategically flooding the web with Chinese-language data. The goal isn't just content creation; it's statistical saturation.

07.12.2025 14:05 πŸ‘ 0 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0

My working theory: We are seeing the early signs of a "Data Offensive."

07.12.2025 14:05 πŸ‘ 0 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0

Has anyone else noticed an increase in random Chinese characters appearing in ChatGPT outputs or hallucinations?

At first glance, it looks like a glitch. But I suspect we are witnessing something much more calculated.

07.12.2025 14:05 πŸ‘ 0 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0

Here is a tour of where my brain went with this: satpugnet.substack.com/p/quiet-echoes

02.12.2025 17:55 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

I wrote briefly about it, but the idea kept nagging at me until I had to write a full post about it.

I have been toying with the concept of character propagation; how our personality is analogous to an echo, and the weird complexity of how that signal ripples through the matrix of humans around us.

02.12.2025 17:55 πŸ‘ 1 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0
Preview
The World is a Fractal A mental model to navigate the endless depth of knowledge

blog.satpugnet.com/p/the-world...

25.06.2025 16:43 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
Post image

New post: The World is a Fractal

A mental model to navigate the endless depth of knowledge

Ever googled something simple, and emerged 3 hours later, 27 tabs deep, questioning reality itself? I explore why that happens.

Featuring spellbinding fractal images.

Link below πŸ‘‡

25.06.2025 16:43 πŸ‘ 1 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0
What is Intelligence? | Antikythera

Truly great read if you like science and new perspective on ideas: whatisintelligence.antikythera.org/

23.06.2025 16:00 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
AI Safety Playlist A curated selection of the best video content on AI safety.

Amazing AI safety playlist to learn about the field: youtube.com/playlist?li...

20.06.2025 16:00 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

Survivor bias, philosophy edition: We inherit ideas only from thinkers who both wanted to broadcast and had the clout to get amplified. Every β€œclassic” bears gatekeepers’ fingerprints; scrutinize them as hard as the arguments.

05.05.2025 16:00 πŸ‘ 1 πŸ” 0 πŸ’¬ 2 πŸ“Œ 0
Post image

Finally! My first brilliant move

500th times is the charm

I am now 0.003 % @MagnusCarlsen by volume

28.04.2025 16:00 πŸ‘ 2 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0