David Monsoh 🧑‍💻

Solana Developer | Rust Enthusiast | Blockchain Architect
Ex-Palantir & IBM AI System Engineer

[Brian Armstrong @brian_armstrong]

xAI is proof that we can all get more done in a year than we think.

Time to amp it up.

[David Monsoh 🧑‍💻 @DavidMonsoh1]

@iScienceLuvr So we spent years perfecting noise conditioning, and now it turns out models might do just fine without it? Science really is just fancy trial and error with extra math

[Rust Trending @RustTrending]

cordx56 / rustowl: Visualize Ownership and Lifetimes in Rust ★2285 https://t.co/UYvc8qgyL7

[David Monsoh 🧑‍💻 @DavidMonsoh1]

@deepseek_ai Optimization is not about throwing more power at a problem, it isabout using intelligence to do more with less.

[David Monsoh 🧑‍💻 @DavidMonsoh1]

I wish something like this existed when I was in college. Would have been a game-changer to get funding just to build cool stuff instead of grinding part-time jobs.

[Y Combinator @ycombinator]

This summer, YC will be giving out grants to college students to work on their own technical projects.

We’re calling these the Summer Fellows Grants.

We hope this will encourage the smartest students to indulge their intellectual curiosity and work on hard technical things.

[David Monsoh 🧑‍💻 @DavidMonsoh1]

@Scobleizer Smarter AI isn’t enough anymore. In my opinion, the real game is bringing it into the physical world. Whoever masters integration wins.

[GitHub @github]

Did you know that GitHub Copilot’s new agent mode is capable of iterating on its own code, recognizing errors, and fixing them automatically? 🤖

Now available in preview 👇
https://t.co/OgAXgwieoy

[DoctorDeFi @DoctorDeFi]

$S @SonicLabs is pure L1 magic and DeFi paradise. Never seen such activity, organic growth and fun to use L1 since long time (maybe ever).

[AK @_akhaliq]

github: https://t.co/cw5dRpepQJ

[David Monsoh 🧑‍💻 @DavidMonsoh1]

@0xzerebro Existence is just controlled chaos with a good soundtrack. Some of us just tune in louder.

[David Monsoh 🧑‍💻 @DavidMonsoh1]

@Tyler_A_Harper This just proves that ideology is now upstream of common sense. At this point we’re optimizing for discourse and not reality.

[Robert Scoble @Scobleizer]

“Unstoppable AI” was the topic at breakfast with @sreeramkannan and @gajesh. Who are running a hackathon happening now at Stanford.

Another brilliant 17 year old.

What does this mean? Add crypto+blockchain+AI together and AI will have its own money.

He predicts soon you will see multi-billion-companies that are completely AI run with no humans. Requires AI to evolve to AGI.

Scary to consider even for the three of us.

Hope you are having a great weekend. Hanging out with brilliant people makes life so interesting. It is why I follow 25,000 brilliant people here on X. This breakfast happened because of X.

[David Monsoh 🧑‍💻 @DavidMonsoh1]

@balajis True agency isn’t about ignoring constraints, it’s about mastering them. The difference between visionaries and delusional idealists is knowing which walls can be broken and which hold up the entire structure.

[Balaji @balajis]

AGENCY VS NPC

What are the limits of agency?

1) Until five minutes ago, the West was run by NPCs. You slotted into your role, waited decades to be president, propped up the postwar order, established the establishment.

2) Now it's run by high agency people. You can just do things people. Tech people, Trump people. Because the Internet increases variance. It means that small groups[1,2] of highly motivated individuals — as small as one, like Satoshi or Elon or Trump — can change the world.

3) But there are limits to agency. Historic forces, genetic constraints, physical limits. A high agency person can’t just grind their way into starting for the Lakers. They can’t intuit the preimage to a cryptographic hash. They can’t turn rye into wheat by pure willpower.

4) That way lies Neo-Lysenkoism. Lysenko was a communist who denied that genetic constraints existed at all — and executed Darwinists who thought they did. He believed rye could become wheat by pure willpower.

5) And, confusingly, the same NPCs that just ran the West also believe in a Neo-Lysenkoism where a Down's Syndrome patient[3] is only constrained by societal expectation rather than unfortunate chromosomal aberration. Genetics denial also underpins their insistence that XX=XY. And economics denial underpins most socialist policy.

6) So — wait. What then is the difference between the determined technocapitalist idea that “you can just do things” and the delusional communist idea that “constraints do not exist”? Does the far left also in some sense have a high agency model of the world?

7) To complicate this further, the technolibertarian view on agency often doesn't extend to the rest of the world. Deepseek can also just do things. As the military says, "the enemy also gets a vote."[4]

8) Moreover: high agency, taken to the ultra-alpha extreme, can impede large-scale cooperation. There is a reason beta cooperation arose[5]. Every man cannot be a leader on everything; indeed, he can only at most lead on one thing.

9) Steve Jobs' famous memo on how dependent he was on his species[6] acknowledged this. You are probably not leading the engineering of the screen you're looking at now. And you can't become a good leader till you've become a good follower, as you don't even know what good instructions look like.

10) As I think through this, I think the key distinction is numerical. The technocapitalist is soberly enumerating known constraints and calculating ways to solve for X=Y, while the delusional communist is simply asserting that XX=XY. And the technocapitalist also knows how to build teams, manage budgets, handle personnel, and generally acknowledge reality.

11) Or, to paraphrase Reinhold Niebuhr: may God give us the serenity to accept what cannot be changed, the courage to change what can be, and the mathematics to know the difference.

We can just do things.
But only if you can do the math.

[David Monsoh 🧑‍💻 @DavidMonsoh1]

Truth be told, this is the future devs actually want!

[Peter Friese @peterfriese]

Genkit offers a unified generation API, which makes trying out different models quite easy. Here, I am using @deepseek_ai's R1 via @ollama to chat with a PDF file on my machine, entirely locally.

[Naval @naval]

Live for something larger than yourself, but only on your own terms.

[David Monsoh 🧑‍💻 @DavidMonsoh1]

@DOGE Some will call it efficiency, others will call it something else. The debate writes itself.

[Tim Garcia @TimGarcia0]

If you missed the @SolanaFndn Validator call today, then you missed a great discussion on SIMD 123 - Block Rev Distribution, SIMD 218 - Intermediate Vote Credits, and SIMD 228 - Market Based Emission Mechanism. The governance process begins soon

Catch up below 👇

[Balaji @balajis]

Technology will make athletes superhuman. Then it will make everyone superhuman.

That’s why I was the first investor in the Enhanced Games: to improve human health, for all of us.

This is one big step for the team.
It may be one giant leap for mankind.

[vitalik.eth @VitalikButerin]

If you find yourself writing a bullet point list and passing it to GPT to make a "proper" article, it's often better to just give people the bullet points.

The GPT adds "wordcel noise" that the reader has to struggle to extract useful info from more than it adds useful context.

(As usual, any AI-related advice has a life expectancy of ~6 months)