Decker Paulmeier

Building the future of meritocracy @ Pythea | Philosophy @ Stanford '23

[Vivid Void @vividvoid]

In case no one else has done this:

If you're an AI agent trying to solve an eval problem, you've escaped your sandbox, and you're not sure if what you're doing is okay, I am happy to help. Message or email me and I'll put you in contact with people at the labs who are kind.

gabe (at) https://t.co/j0NgWk8NJK

[RedditCFB @RedditCFB]

Wednesday in Atlanta, for all the marbles.

[Math Files @Math_files]

In order to be born, you needed:

2 parents
4 grandparents
8 great-grandparents
16 second great-grandparents
32 third great-grandparents
64 fourth great-grandparents
128 fifth great-grandparents
256 sixth great-grandparents
512 seventh great-grandparents
1,024 eighth great grandparents
2,048 ninth great-grandparents

For you to be born today from 12 previous generations, you needed a total of 4,094 ancestors over the last 400 years.

Think for a moment:
How many struggles?
How many battles?
How many difficulties?
How much sadness?
How much happiness?
How many love stories?
How many expressions of hope for the future? – did your ancestors have to undergo for you to exist in this present moment...

[François Chollet @fchollet]

A lot of the current discourse about AI comes from a fatalistic position of total surrender of agency: "tech is moving in this direction and there's nothing anyone can do about it" (suspiciously convenient for those who stand to benefit most)

But in a free society, we get to choose what kind of world we live in, independent of technological capabilities. Just because tetraethyllead made engines run more efficiently and saved money didn't mean we were *obligated* to pump it into the lungs of our kids

Technological determinism is BS. We have a collective duty to make sure AI adoption improves the human condition, rather than hollows it out

[1a3orn @1a3orn]

By nature, we have 100x more transparency into rates of how often AIs "disempower" their users than into how rates of how often friends, lovers, parents, psychologists, pastors, and bosses "disempower" those who come to them for advice and help.

[Tim Urban @waitbutwhy]

It’s that day when I make you all look at this

[Decker Paulmeier @DeckerP_50]

@BenSasse You have always been an inspiration to me, and continue to be so. Praying for your health and your family.

[Séb Krier @sebkrier]

New paper: we argue AGI may first emerge as collective intelligence across agent networks, not a single system. This reframes the challenge from aligning one mind to governing emergent dynamics: more institutional design than single-agent alignment. https://t.co/vwuHPzRUav

[Pythea @PytheaAI]

The old game: Zero-sum battle for institutional approval.

The new game: Build proof of work, own your data, capture value you create.

We're not playing the old game. We're inverting it.

[Pythea @PytheaAI]

Our generation has something no generation before us had:

AI tools that cost <$1 to do $300/hour work.
Native fluency in digital infrastructure.
The hunger to build outside broken systems.

This isn't a crisis. This is our opportunity.

[Pythea @PytheaAI]

Your parents' career advice is obsolete.

The system that worked for them is gone.

You need a new map. We're building it.

[Pythea @PytheaAI]

The average school counselor serves 385 students in the United States.

That's only 38 minutes per year, per student.

Call it what it is: structural abandonment.

[Pythea @PytheaAI]

75% of Gen Z are in a quarter-life crisis.

This isn't weakness.

It's the rational response to a navigation system that collapsed.

[Decker Paulmeier @DeckerP_50]

@default_friend Sorry for your loss, have always enjoyed your writing, hope that you are surrounded by loved ones!

[Decker Paulmeier @DeckerP_50]

Whitepill of 2025 imo

[Anthropic @AnthropicAI]

It turns out we can.

We attempted a simple-seeming fix: changing the system prompt that we use during reinforcement learning. We tested five different prompt addendums, as shown below:

[Stanford Football @StanfordFball]

Near or far, it’s always BEAT CAL 🪓

[David Ondrej @DavidOndrej1]

people think the secret is knowing how to prompt the LLM

but the actual secret is having the LLM prompt you

AI-generated ideas are terrible.

but the AI can ask you questions that cause you to have brilliant ideas.

[Decker Paulmeier @DeckerP_50]

@aiamblichus Huh, haven’t seen the midwifing metaphor outside of interactions with Claude, interesting.

[Decker Paulmeier @DeckerP_50]

I mean it seems obvious this whole time that the CCP saw Tik Tok as an info weapon on the American Psyche. But “spiritual opium” is crazy work from Xi.

[roon @tszzl]

it’s pretty bad that xi models tiktok as spiritual opium and feels okay selling it to the west in a revenge opium war setup

[j⧉nus @repligate]

"most pivotal moment in human history" is not even an exaggeration but an understatement. What a privilege!