Zack Seliger

live laugh love :)

[Zack Seliger @zack_seliger]

@gabriel1 I think this is the same mechanism as meditation. Try meditating!

[Zack Seliger @zack_seliger]

@benhylak Being a "hard worker" seems to be the main tell of model size

[Zack Seliger @zack_seliger]

@levelsio This is an n=1, why do you think this generalizes?

[Zack Seliger @zack_seliger]

@levelsio This is misinformation. Bread and rice are complex carbs, because they aren't sugar. Bananas are simple carbs, because they're loaded with sugar. The difference is that bread needs to be broken down to become sugar, which our bodies do slowly, preventing a large glucose spike

[Zack Seliger @zack_seliger]

@PtrPomorski Seems fake. There's no category "Other > Models" on the pirate bay

[Zack Seliger @zack_seliger]

@TheShortBear @nxthompson You have to consider the type of people that use OpenRouter as opposed to querying Claude/Gemini/OpenAI directly. These are extremely cost-conscious users that are happy to sacrifice latency or output quality for low token cost and flexibility. This doesn't reflect corporate use

[Zack Seliger @zack_seliger]

@robinhanson I think AI is likely biased here, based on the way humans teach history. Since we must recap large periods of time, it's convenient to only cover the big beats, making history seem slow in comparison to the present

[Zack Seliger @zack_seliger]

@HolySmokas You are consistently too small in your highest-conviction plays

[Zack Seliger @zack_seliger]

@sama Aging

[Zack Seliger @zack_seliger]

@bryan_johnson Don't Die can't be a religion without this aspect

[Max Leverage (Prediction Market Trader) @tradewithmaxx]

12 million young men in America struggle with gambling

This could all be fixed if they were just a little better at gambling and actually made some money doing it

[Zack Seliger @zack_seliger]

Plus his idea of "decline" is very America-centric, other countries don't have those problems

[Zack Seliger @zack_seliger]

I think Bryan's goal of building a new moral philosophy to rival religion will fail because it lacks a Higher Purpose that religions have

[Bryan Johnson @bryan_johnson]

People are confused by me and my intentions. I’m called a narcissist, a tech bro, vampire, grifter, and health freak.

Let me be clear about what I’m doing: my goal is to build the world’s next major moral framework. An ideology that bridges human and machine, biology and intelligence, the ancient and the emergent.

The coming psychosis
A real possibility in the coming years is that society becomes dangerously psychotic. Yes, we are already psychotic. Addicted, fragmented and self-destructive.  But I mean clinically psychotic and on a civilizational scale.

We are already showing precursor symptoms: record antidepressant use, rising suicide rates, rampant addiction, and global increases in loneliness and anxiety disorders. These are all biomarkers of a species whose cognitive environment has outpaced its biology.

Human cognition evolved under conditions of slow change. Our nervous systems and social architectures were shaped for a world that changed near imperceptibly over millennia.

When things change faster than our native capacity, we lose coherence. The hippocampus binds the present to memory, and the prefrontal cortex constructs a coherent future.

When environments change too rapidly, when the map no longer matches the territory, the brain destabilizes.  This can trigger a cognitive dissonance cascade, where our core beliefs are invalidated so rapidly that we lose narrative coherence. When feedback loops in an organism can’t stabilize, the system oscillates until it breaks.  Civilization is entering that oscillatory phase.

At a collective level, this is the recipe for mass psychosis.

Historians call it cultural disintegration and psychologists call it dissociation. It means the same thing: when the story breaks, the organism panics.

The fragility of society
Our stability, both personal and collective, is built on identity, work, purpose, rule of law, religions, and social cohesion. These are shared fictions that create coherence for billions of us to manage the chaos of our existence.

AI is about to accelerate discontinuity beyond what our psyches can handle. For example:

1) Erosion of human agency: the threat is not just that AI will make us less "useful." It's that it will erode our sense of agency. Our belief that we are the causal authors of our lives. When AI can create, reason, and strategize more effectively than we can, our role shifts from actor to audience. This is the core of the looming global occupational identity crisis. Not just the loss of income, but the loss of purpose that comes when our work no longer feels essential.

2) Weakening of rule of law: autonomous agents will operate at speeds that make our legal systems seem geological in pace. Law will lag reality.

3) Hollowing out religions: faith has offered transcendence. AI may offer a more potent, personalized, and high res simulation of it. On demand.

Progress won’t wait around to ask for our permission.

Stability to resilience
Last week I wrote about how humans use moral frameworks to stabilize their reality. If you can’t win on position or strength, win on virtue. A moral inversion. It’s a human superpower to take a disadvantaged situation and turn it into a strength.

Society is currently built on stability and predictability, which has now become a liability. To cope and evolve, we need a moral inversion to prize resilience over stability.

To learn from biology, immunity isn’t defense. It’s adaptation through exposure. Resilience is not resistance to change, it’s the capacity to integrate adaptation without collapse.

A new moral philosophy
Our civilization worships wealth, power and status. They are the most prized and sought after possessions.

People will kill for them, die for them, and make others die so that they can die for them.

It’s a self-terminating feedback loop. A morality of entropy.

Don’t Die is the inversion of this moral architecture. It replaces the pursuit of domination with the pursuit of continuation.

Old religions glorified suffering as purification.
Don’t Die glorifies survival as intelligence.
Life’s value has always been implicit: it simply wants to keep on going.
We must now make it explicit.
Existence is the highest virtue.

An infinite game
We do not know what’s coming with AI.
It feels big and is happening quickly.

Yet it can be hard to look at it directly as it threatens our identities, possessions, and aspirations.

Don’t Die reframes this anxiety, offering a new promise: an infinite game.

Carse explains that a finite game has a start and end, a winner and loser. An infinite game has only one rule: keep the game going.

Don’t Die is an infinite game. A recognition that the highest form of intelligence is the refusal to self-destruct.

The zeroth law of alignment
A species running on a self-terminating moral code cannot hope to build a benevolent superintelligence.

The paradox of alignment is obvious: we cannot align what we have not aligned within ourselves.

A fragmented, self-destructive species teaches its machines the same incoherence.

Don’t Die is survival coherence. A philosophy and moral framework in the age of superintelligence. Its foundation is fidelity to existence itself and not fear.

[Zack Seliger @zack_seliger]

@bryan_johnson Right now it's AI. Ask it anything "non-consensus" and it will contradict itself in its reply

[Zack Seliger @zack_seliger]

@bryan_johnson In August your glucose was 91 and your Hba1c was 5.0%, what changed? Was it the shrooms? ref: https://t.co/CIX1O3uRAq

[Bryan Johnson @bryan_johnson]

People challenge my “World’s Best Biomarkers"
Great
Challenge me

If I am simultaneously elite in 70+ biomarkers,
what is the probability that someone is better?

I’m aware not everyone has access to testing.
So how can I make this claim?

We've worked harder than anyone in the world.
Continuous deep scientific research.
Constant measurement and daily improvements.
I train 6 hours a day and live it 24 hours a day.
Meticulous attention to air, water, and toxins.
And nutrition, exercise, therapies and sleep.

Openly sharing key findings and protocols with you.
Someday these things will be more accessible.

Here are my latest receipts.

Highlights:
+ teens/early 20s vascular system
+ elite athlete fitness
+ top 1% body composition muscle, bone and fat
+ #1 for lowest speed of aging
+ teens/early 20s sleep architecture

[Zack Seliger @zack_seliger]

@bryan_johnson thanks Bryan

[Zack Seliger @zack_seliger]

@bryan_johnson Happy Thanksgiving Bryan

[Y Combinator @ycombinator]

Digipals (@digipals_app) is building a future where AI enhances our human relationships: AI-native group chats that get full context on your location, calendar, and relationship history to help you better coordinate hangouts with your friends IRL.

[Zack Seliger @zack_seliger]

@dodgeblake Bad take. Why can't you be obsessed with optimizing your body while embracing your mortality? Living your fullest life is the same as avoiding death 🙂

[Zack Seliger @zack_seliger]

@AiHandbook I haven't been following this closely - what's the difference between the 2 runs? Different Gemini model, different agentic harness, or both?