Kevin Schmidt @kevinschmidt.bsky.social

ML at Instagram, former CTO @ThisIsCentury, former head data cruncher at @MindCandy and @lastfm, speaker @strataconf/@nuclai, @kevinschmidt.bsky.social

[Nick St. Pierre @nickfloats]

I ran all of the Sora prompts through Midjourney

Interesting how similar some are

side-by-sides against vids:

[John Burn-Murdoch @jburnmurdoch]

NEW: an ideological divide is emerging between young men and women in many countries around the world.

I think this one of the most important social trends unfolding today, and provides the answer to several puzzles.

[Gary Marcus @GaryMarcus]

How much was OpenAI really worth, a week ago? And what are the employees thinking about right now?

Here’s a thought:

If OpenAI has a lot of breakthrough, difficult to replicate IP (code, data, infrastructure etc), with a real business case, a lot of employees would stick around, rather than giving up the unvested part of their stake in something massive.

If it was never really worth what people thought, and everything could easily be replicated, everybody will bolt.

If they do, it’s a strong indication that generative AI has been overvalued.

[Gary Marcus @GaryMarcus]

Terrifying and correct thought.

Three orgs filled with brilliant minds tried to create AI independently of the tech giants, and all three have been subverted.

[François Chollet @fchollet]

Turns out, fitting a curve to a dataset produces a model that only generalizes to that specific data distribution -- big if true.

[François Chollet @fchollet]

"Calculators have superhuman math abilities -- both mathematicians and accountants will soon be out of a job"

[Filip Piekniewski🌻 🐘:@[email protected] @filippie509]

Updated (yet again) state of autonomous vehicle scene late 2023:

Starsky ☠️
Argo ☠️
Uber/Otto ☠️
Embark ☠️
Locomation ☠️
Optimus Ride ☠️
TuSimple 🧟
Nuro 🧟
Cruise 🧟
Tesla 🤡
Waymo 💸
Aurora 💸

[Gary Marcus @GaryMarcus]

Disappointed by the new “AGI is here” @blaiseaguera @NorvigPeter article. it starts with a great analogy to ENIAC but the logic is flawed, as explained below 🧵

[François Chollet @fchollet]

My interpretation of prompt engineering is this:

1. A LLM is a repository of many (millions) of vector programs mined from human-generated data, learned implicitly as a by-product of language compression. A "vector program" is just a very non-linear function that maps part of the latent space unto itself.

2. When you're prompting, you're fetching one of these programs and running it on an input -- part of your prompt serves as a kind of "program key" (as in database key) and part serves as program argument(s). Like, in "write this paragraph in the style of Shakespeare: {my paragraph}", the part "write this paragraph in the stye of X: Y" is a program key, with arguments X=Shakespeare and Y={my paragraph}.

3. The program fetched by your key may or may not work well for the task at hand. There's no reason why it should be optimal. There are lots of related programs to choose from.

4. Prompt engineering represents a search over many keys in order a find a program that is empirically more accurate for what you're trying to do. It's no different than trying different keywords when searching for a Python library.

5. Everything else is unnecessary anthropomorphism on the part of the prompter. You're not talking to a human who understands language the way you do. Stop pretending you are.

[NYTPitchbot @DougJBalloon]

Whether it’s critics saying the 14th amendment bars Trump from running for president because of seditious activity or supporters saying the 22nd amendment bars him from running because he has already won two presidential elections, both sides say Trump is ineligible in 2024

[Noah Smith 🐇🇺🇸🇺🇦🇹🇼 @Noahpinion]

Remember two years ago when the CCP was this infinitely competent juggernaut that everyone thought controlled the entire Chinese economy down to the tiniest details and did everything optimally and blah blah blah?

[warm human @w000rm]

spring:
depressed in a hopeful way

summer:
depressed in a manic way

fall:
depressed in a nostalgic way

winter:
depressed in a depressed way

[Max Woolf @minimaxir]

The rare self-own due to data science

[François Chollet @fchollet]

Big difference between memorizing reusable programs (thus demonstrating skill at one known task) vs. synthesizing new programs on the fly (thus being able to adapt to arbitrary new tasks that cannot be anticipated in advance)

[Cory Doctorow NO LONGER ON TWIT TER @doctorow]

Elon Musk lies a lot. He lies about being a "utopian socialist." He lies about being a "free speech absolutist." He lies about which companies he founded:

https://t.co/2n1OvrKQNJ

1/

[Shit User Story @ShitUserStory]

As a…
– Twitter user
I want…
– unlimited interactivity – centered in audio, video, messaging, payments/banking – creating a global marketplace for ideas, goods, services, and opportunities.
so that…
– It’s no longer possible to even parody this fucking shitshow.

[Wild Geerters @steinkobbe]

There were some incredible posts at the height of the NFT bubble

[LowPassFilter @hemant_4771]

In 2017, I was working on driverless car at IITB. I can confirm that this is 100% true. Almost everyone underestimated the complexity of problem & how far the tech is in reaching level 5 autonomy. So yeah, stop living in la la land of hype cycles and have reasonable error bounds

[Classical Studies Memes @CSMFHT]

Don't make me tap the sign