←
Jace Kohlmeier1,016 posts
Jace Kohlmeier avatar

Jace Kohlmeier

@derandomized

Entrepreneur, data scientist, investor, and family man. // Former data science at Khan Academy, co-founder at Teza Technologies, Citadel HFT.

◷ Joined March 2012▧ 37 Photos and videos
roon avatar
⟳ Jace Kohlmeier retweeted
roon@tszzl·14h

yes we can neither conclude nor rule out an intelligence explosion from this IMO the real tell will be when models' research taste exceeds the research taste of good researchers if there is a constant or increasing slope. i would certainly expect it to be higher multiple than this 2x x.com/mert_gulsun/status/212…

There is a chance we will see constant acceleration in the form of a series of successive trend breaks

Sam Pullara avatar
⟳ Jace Kohlmeier retweeted
Sam Pullara@sampullara·Nov 13

claude, rewrite the FAA in Rust, make no mistakes

The @FAAnews issued a nationwide ground stop for all U.S. airports following a communication issue impacting all Air Route Traffic Control Centers. FAA technicians are investigating the issue. We will provide immediate updates as we have them.

Jace Kohlmeier avatar
Jace Kohlmeier@derandomized·Nov 14

Inside me are two wolves. Tomorrow around lunch I’m gonna feed a bunch of delicious snacks to the rocket watching wolf.

There are two types of people: those glued to their screens watching FAA airline closures, boats running into each other, and CEO infidelity scandals. And those glued to their screens watching the world’s first orbital class rocket be caught on a tour after circling the globe 22 times.

Patrick Collison avatar
⟳ Jace Kohlmeier retweeted
Patrick Collison@patrickc·Nov 8

Two things I’ve been thinking about regarding AI: • People often highlight the risks involved in AI, which are surely very real. I’m struck, however, by how hard it is to understand the contours of the positive scenario. There are easy-to-perceive and frequently invoked elements of a positive vision — more diseases cured, more material abundance, etc. — which sound great, but these are by themselves just ingredients. I’m curious what the larger equilibrium is supposed to look like. Do people work? If so, why? What is their marginal contribution? If not, how are resources allocated? What is scarce? What is the ratio between humans and AIs? Are humans in control, whatever that means? One sometimes hears the Culture novels invoked as one scenario, but I’m not sure they depict a desirable outcome. (AI runs everything and humans compete in games? Iain Banks himself said that the humans in the Culture are “somewhere between passengers, pets and parasites.”) There is Richard Brautigan’s famous poem, but do we want to be “returned to our mammal brothers and sisters”? (Isn’t that rewinding Genesis?) We could say that it was hard to imagine the world of 1900 from the vantage point of 1700, and that’s assuredly true. But nobody in 1700 was trying to recast humanity itself within a couple of years — things necessarily unfolded slowly and by increments, and therefore with the possibility for error correction. Given how rapid the rate of progress in AI may be, it feels strange for the upside case to be so opaque. • There is increasing discussion of whether AIs are conscious or how we would know whether they are, with conclusions there often assumed to be relevant to matters of moral worth. At first blush, this seems a rather arcane, benign, or even absurd issue. (Remember the response to Blake Lemoine’s fears.) Setting aside the terms of the debate itself, or the relevant demarcation criteria, it seems to me a quite high-stakes question, for an affirmative answer — that is, granting moral worth to AIs — would presumably mean that the aggregate worth of AIs would soon outweigh that of humans in any utilitarian calculus. The AIs will be exceptionally wise and intelligent, and vastly more numerous (especially once one considers the imminent future and the AIs not yet born). As such, even epsilon of moral worth will multiply to a lot once one considers the quadrillions that could soon exist. (This is effectively a distributed version of the Nozick’s utility monster.) If this conclusion is unpalatable, I think one of three propositions needs to be rejected: the premise AIs can have moral worth; utilitarianism itself; or the belief that there is even a possibility that they will be far more numerous than humans. With both of these matters (contours of future society; AI moral worth), there is the possibility of dramatic change. Today, people argue about datacenters. I’m curious how popular support for AI will evolve as more people come to understand these medium-term questions and the relevant stakes. What is the vision for AI that most people would be excited about?

Brandon Beylo avatar
⟳ Jace Kohlmeier retweeted
Brandon Beylo@marketplunger1·Nov 5

Hearing unconfirmed rumors that my wife has no idea what's happening in markets and she's living the most blissful, cheerful life possible. Inspiring.

tae kim avatar
⟳ Jace Kohlmeier retweeted
tae kim@firstadopter·Nov 4

Gina is a politician. She cares about her constituents. She cares about employment and guaranteeing a lot of union jobs. She cares about winning concessions such as no subsidies and promises on water and electricity supply / costs. Those will be big public wins for her. This is a solvable issue. It's time for AI companies to get to work.

Re: Texas governor election. Let's take a deep breath. If companies ask for no subsidies, promise a ton of union jobs, and guarantee that water and electricity costs won't rise/plus add supply, data centers can be built under her governorship. This is directly from Gina's website:

Post attachment
Jace Kohlmeier avatar
Jace Kohlmeier@derandomized·Oct 8

👀

I did tell @tylercowen that if forced to pick a market price as indicator of the singularity, it would be sovereign debt yields. If the “gradual disempowerment” future is real, its way will likely be paved by very high interest rates.

Childless Garbanzo avatar
⟳ Jace Kohlmeier retweeted
Childless Garbanzo@MoLeiTowel·Sep 17

LLM that’s trained solely on Modest Mouse, Death Cab for Cutie, and The National and becomes too withdrawn and awkward to take over the world

Jace Kohlmeier avatar
Jace Kohlmeier@derandomized·Sep 14

While I'd quibble with a few details of this presentation-- better to use Joules per completed task--batching and speed noramalizations are still enormous updates to my mental model. I've been reading these power comparisons for a long time. Kinda embarassed that I missed this.

This number is way off. You’re comparing one human brain to an entire data center capable of training new models. The actual answer is: IQ points per watt: Humans: 5 AI: 7 - 40 (!!) By my rough calculations the current AI is already served more efficiently than humans for equivalently intelligent work performed! A single gpu can produce outputs and parse inputs way faster than a human and can do many of such requests in parallel. A larger model generally mostly means you need more gpus to serve it, but that often also increases how many requests you can serve in parallel, so I’m reporting on the share of the serving system, and also normalizing for speed. Really a better way to measure and report it is joules per equally good completed task, as your formulation suggests we can likely add more watts to get more iq, which isn’t correct. I did some back of the napkin calculations based on the latest serving hardware and the most powerful open source models. But even given pretty generous assumptions it turns outout computers come ahead. For example the request is allocated approximately 60 W while running, but finishes a task in about one-seventh of the human time.

Jay Kreps avatar
⟳ Jace Kohlmeier retweeted
Jay Kreps@jaykreps·Sep 12

It's worth pointing out that practically speaking, the alternative to a sane regulatory regime isn't "no regulation." We are experiencing rapid improvements in capabilities, have multiple providers with differing levels of safe guards, and are at the point now where small things are starting to go wrong. The scope of AI capabilities are growing rapidly, and the scope of possible damage is going to grow with them. Think for a second about what happens when the first actually bad and "politically gripping" incident occurs? Not Hugging Face, but something that makes the average person really scared. This isn't about extinction, think a human initiated, AI-orchestrated cyberattack that sabotages one piece of infrastructure that results in at least one person dying. The reaction will be violent. Our political system doesn't work well at proactively handling purely theoretical problems, but reacts very fiercely and sometimes irrationally when something bad actually happens. The August 2001 CIA memo "Bin Laden Determined to Strike in US" got little response, but buildings falling down changed the course of history. An industry whose wealth and power is widely resented, that resists even the most innocuous and common sense safeguards, and then causes a significant incident, will see a violent reaction. This will result in a regulatory regime designed in a moment of populist outrage by people with little understanding of what will help, whose primary goal will be crushing AI progress as an end in its own right.

We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so. Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training. You can read the full post here: darioamodei.com/post/we-must…

Jace Kohlmeier avatar
Jace Kohlmeier@derandomized·Sep 12

Extremely useful thinking: Dichotomies of bounded and unbounded, frontier and cheap; Reflexivity specific to unbounded. It should be everyone’s hope that the “unbounded” category revenue diversifies -> de-correlates -> derisks. Biopharma, materials science, etc.

Giovanni Cattani @giovannicatt3 Sep 1 Article Nobody is talking seriously about AI demand AI capex for 2028 is forecast to be larger than the budget of France. Frontier AI labs’ revenue ramp justifies almost any number. Reflexivity in AI demand is a double-edged sword, and it's now a good

College Sports Only avatar
⟳ Jace Kohlmeier retweeted
College Sports Only@CSOonX·Sep 11

Todd Beamer was the passenger on United Airlines Flight 93 who rallied fellow passengers to fight back against the 9/11 hijackers with the phrase "Let's roll" and stopped it from reaching Washington, DC. This is his daughter, Morgan, who was born 4 months after his death and went on to play women's soccer at his alma mater, Wheaton College. Imagine how proud he would be. 🇺🇸 #neverforget

Post attachment
Clay Dumas avatar
⟳ Jace Kohlmeier retweeted
Clay Dumas@claydumas·Sep 9

This strategy won't pan out for hard tech. Investors need to care just as much as founders that more water actually falls out of clouds, electrons flow out of a power plant, lithium comes out of the brine, energy is stored in the battery, etc. The industrial tourists who would be just as fulfilled bidding up cartoon monkey JPEGs will not have the stomach for the ups and downs of building real-world companies, and founders building civilizationally important technology are going to get left stranded.

Had a growth investor at a very well respected VC firm tell me (in all sincerity and honesty), “My job isn’t to find great companies, it’s to find companies that will be marked up. TBH, what happens to these companies long-term doesn’t really matter.”