Adam Klimont

PGP: 703474FA

[Tim Urban @waitbutwhy]

Good litmus test for which areas of personal growth to focus on:

Ask yourself, would you want your child to marry someone just like you?

If the answer is "yes, except ABC," then A B and C are what you should work on. (If the answer is "no" then there's even more work to do.)

[Michael Nielsen @michael_nielsen]

A rule I've found helpful: if invited to something, especially professional (e.g., give a talk, attend a workshop, etc), ask myself "If this were tomorrow, would I be feeling anticipation & excitement, or be feeling a bit annoyed with myself for accepting?"

[norvid_studies @norvid_studies]

thread of stupid questions

[Kevin Fischer @kevinafischer]

I don't talk much about this - I obtained one of the first FDA approvals in ML + radiology and it informs much of how I think about AI systems and their impact on the world. If you're a pure technologist, you should read the following:

There's so much to unpack for both why Geoff was wrong, and why his future predictions should not be taken seriously either.

Geoff made a classic error that technologists often make, which is to observe a particular behavior (identifying some subset of radiology scans correctly) against some task (identifying hemorrhage on CT head scans correctly), and then to extrapolate based on that task alone.

The reality is that reducing any job, especially a wildly complex job that requires a decade of training, to a handful of tasks is quite absurd.

Here's a bunch of stuff you wouldn't know about radiologists unless you built an AI company WITH them instead of opining about their job disappearing from an ivory tower.

(1) Radiologists are NOT performing 2d pattern recognition - they have a 3d world model of the brain and its physical dynamics in their head. The motion and behavior of their brain to various traumas informs their prediction of hemorrhage determination.

(2) Radiologists have a whole host of grounded models to make determinations, and actually, one of the most important first order determination they make is whether there is anything notably wrong with a brain structure that "feels" off. As a result, classifiers aren’t actually performing the same task even as radiologists.

(3) Radiologists, because they have a grounded brain model, only need to see a single example of a rare and obscure condition to both remember it and identify it in the future. This long tail of rare conditions to avoid missing is a large part of their training, and no one has any clue how to make a model that acts similar in this way.

(4) There’s so many ways to make Radiologist lives easier instead of just replacing them, it doesn’t even make sense to try. I interviewed and hired 25 radiologists, whose primary and chief complaint was that they had to reboot their computers several times a day.

(5) A large part of the radiologist job is communicating their findings with physicians, so if you are thinking about automating them away you also need to understand the complex interactions between them and different clinics, which often are unique.

(6) Every hospital is a snowflake, data is held under lock and key, so your algorithm might not work in a bunch of hospitals. Worse, the imagenet datasets have such wildly different feature sets they don’t do much for pretraining for you.

(7) Have you ever tried to make anything in healthcare? The entire system is optimized to avoid introducing any harm to patients - explaining the ramifications of that would take an entire book, but suffice to say even if you had an algorithm that could automate away radiologists I don’t even know if you could create a viable adoption strategy in the US regulatory environment.

(8) The reality is that for every application, the amount of specific and UNKNOWABLE domain knowledge is immense.

LONG STORY SHORT: thinkers have a pattern where they are so divorced from implementation details that applications seem trivial, when in reality, the small details are exactly where value accrues.

Should you be worried about GPT5 being used to automate vulnerability detection on websites before they’re patched? Maybe.

Should you be worried GPT5 is going to interact with SOCIAL systems and destroy our society single-handedly? No absolutely not.

[Andrej Karpathy @karpathy]

[New Talk] Pleasure to come by Microsoft BUILD this year and give a talk on "State of GPT". Goes through the GPT Assistant training pipeline, covers some "LLM Psychology", and offers a few best practices:

https://t.co/HDJix905Gy

[Paul Graham @paulg]

@engineers_feed Travel before you have kids.

[Dr Charlotte Houldcroft @DrCJ_Houldcroft]

My favourite study of "how often does a family with small children get a respiratory infection?" is this one, from the USA:
https://t.co/fepEzs4A6Q
Children <5y have one of 16 common respiratory viruses detected in a nasal swab 50% of the year!

[じゃがりきん @jagarikin]

あの伝説のルービックキューブをさらにわかりやすくしました

[Fascinating @fasc1nate]

A ball rolling down a massive xylophone plays Bach's "Jesu, Joy of Man's Desiring" in a forest in Japan.

[Jay Alammar @JayAlammar]

AI image generation is the most recent mind-blowing AI capability.

#StableDiffusion is a clear milestone in this development because it made a high-performance model available to the masses.

This is how it works.

1/n

[Adam Klimont @aklimont]

@robhague Have you tried https://t.co/ec9qdiliWJ?

[Erik Brynjolfsson @erikbryn]

Superpower:
More inventors immigrated to America than to all other nations in the world combined between 2000-2010.

https://t.co/O3RfJ8MVmk

HT: @AlecStapp and @JeremyLNeufeld

[Briar Adams @BriarAdams]

Impactful applications of #deeplearning : @bartczakt @aklimont share impressive progress in #medical #imaging #segmentation for improving clinical decision making using #computed #tomography @pydatalondon #healthcare #datascience

[Cydar Medical @CydarMedical]

We are proud to be at the forefront of using #innovation to help deliver better treatments to patients. Watch below where Dr Rachel Clough discuss the trial in more detail.
#augmentedintelligence, #cydarproud, #futuretechnologies

[Elon Musk @elonmusk]

Tap on the stars in upper right of screen to revert to chronological

[Adam Klimont @aklimont]

@dontusethiscode (#DevelopingExpertiseInPythonCourse): "assertions are programmatic documentation"

[Cydar Medical @CydarMedical]

Very excited to be collaborating with Medtronic on this pilot program.
#cydarmedical, #futuretechnologies, #collaboration, #medtronic, #augmentedintelligence, #revolutionizinghealthcare

Learn more here: https://t.co/vge57DemqM

[Adam Klimont @aklimont]

TIL from @dontusethiscode: In Pandas, square brackets by themselves are ambiguous when accessing elements. Use .loc and .iloc to disambiguate

[Adam Klimont @aklimont]

Crossing the boundary of "Restricted Computation Domains" like NumPy often comes with a performance hit. Good example by @dontusethiscode:

[Adam Klimont @aklimont]

Taking a 3-day Python&pandas course by @dontusethiscode. Loving the focus on the design/mental models for the "restricted domain" tools like NumPy and pandas. Looking forward to learning more about async/await tomorrow!