spectate_or's profile picture. ml since 2012. also cs, phil, econ, geopol, law.  phd student on latent spaces of LLMs. normativity is fiction, reject it.

Daniel Kaiser

@spectate_or

ml since 2012. also cs, phil, econ, geopol, law. phd student on latent spaces of LLMs. normativity is fiction, reject it.

Pinned

My new work with @GBR_Data is on Arxiv now. arxiv.org/abs/2509.18458 🧵We introduce a reasoning benchmark for LLMs where you can vary difficulty, length, and noise truly independently. It's also the first benchmark that grounds these dimensions in Cognitive Load Theory.


Daniel Kaiser reposted

RESEARCH: Night owls, or adults who stay up late and have a preference for evening activity, generally performed better on cognitive tests than people who are morning-oriented, according to study by Imperial College.


Maybe “the world failed” but let’s also recognize the UN failed at making the world succeed, and this is not the only issue where it failed. Does the UN, its institutions, and their powers need reform? What point does an entity have that constantly fails crucial goals?

🌎 The world has "failed" in its promise to hold warming below 1.5C, UN Secretary-General Antonio Guterres has told global leaders in Brazil, warning of greater costs and destruction for every nation. At a leaders summit in Belem ahead of COP30 climate talks, Guterres said…

AFP's tweet image. 🌎 The world has "failed" in its promise to hold warming below 1.5C, UN Secretary-General Antonio Guterres has told global leaders in Brazil, warning of greater costs and destruction for every nation.
 
At a leaders summit in Belem ahead of COP30 climate talks, Guterres said…


Daniel Kaiser reposted

ML research is an engineering discipline, not a philosophy seminar. You build, you test, you learn. Untested ideas are just speculation.


Anthropic's retirement plan for older models include an exit interview

krishnanrohit's tweet image. Anthropic's retirement plan for older models include an exit interview


Daniel Kaiser reposted

Continuous Autoregressive Language Models Compresses a fixed number of tokens into a vector using an autoencoder, and directly models prediction of next vector.

iScienceLuvr's tweet image. Continuous Autoregressive Language Models

Compresses a fixed number of tokens into a vector using an autoencoder, and directly models prediction of next vector.

Daniel Kaiser reposted

. @AnthropicAI - how on Earth does this possibly trigger a safety filter?!

peterwildeford's tweet image. . @AnthropicAI - how on Earth does this possibly trigger a safety filter?!

Daniel Kaiser reposted

Hype / reality

liron's tweet image. Hype / reality
liron's tweet image. Hype / reality

Today's Extropic launch raises some new red flags. I started following this company when they refused to explain the input/output spec of what they're building, leaving us waiting to get clarification.) Here are 3 red flags from today: 1. From extropic.ai/writing/inside…

liron's tweet image. Today's Extropic launch raises some new red flags.

I started following this company when they refused to explain the input/output spec of what they're building, leaving us waiting to get clarification.)

Here are 3 red flags from today:

1. From extropic.ai/writing/inside……


if useless shit is inside, cover it with gold on the outside

“setting up my TSU”

stevejang's tweet image. “setting up my TSU”


its sad how the X algorithm went to shit. i used to see the posts of my friends and interesting research but its all gone


Daniel Kaiser reposted

There are two kinds of AI doomers: people who don't know anything about AI and AI researchers who've gotten carried away with their recent progress. What they have in common is that they've both watched too many bad sci-fi movies.


Daniel Kaiser reposted

The bureaucracy is beyond parody. What a waste of taxpayer money- like the city owned grocery stores for NY, these EU owned AI Factories are designed to fail.

🇪🇺 As a European citizen and AI founder, I can apparently use these "AI Factories", so I just signed up to use them! Every "supercomputer" has an [ ACCESS NOW ] button which made me very excited I expected to sign up, maybe pay a discounted H100 rate (funded by EU, that'd be…



Daniel Kaiser reposted

The problem is there's not a single tech person in this entire group It's bureaucrats bureacrating!

levelsio's tweet image. The problem is there's not a single tech person in this entire group

It's bureaucrats bureacrating!

💻 I just wanted to show how easy it is getting a GPU in the regular way compared to the 🇪🇺 EU's "AI Factory" plan where you have to apply for a proposal Funnily enough @LambdaAPI actually shows "Design your AI Factory" on their landing, maybe they're trying to get that juicy EU…



Daniel Kaiser reposted

Vague statements like this, which fundamentally cannot be operationalized in policy but feel nice to sign, are counterproductive and silly. Just as they were two or so years ago, when we went through another cycle of nebulous AI-statement-signing. Let’s set aside the total lack…

deanwball's tweet image. Vague statements like this, which fundamentally cannot be operationalized in policy but feel nice to sign, are counterproductive and silly. Just as they were two or so years ago, when we went through another cycle of nebulous AI-statement-signing.

Let’s set aside the total lack…

A stunningly broad coalition has come out against Skynet: AI researchers, faith leaders, business pioneers, policymakers, NatSec folks and actors stand together, from Bannon & Beck to Hinton, Wozniak & Prince Harry. We stand together because we want a human future.…

tegmark's tweet image. A stunningly broad coalition has come out against Skynet: AI researchers, faith leaders, business pioneers, policymakers, NatSec folks and actors stand together, from Bannon & Beck to Hinton, Wozniak & Prince Harry.  We stand together because we want a human future.…


spectate_or's tweet image.

Congratulations @jpmorgan on the opening of your new headquarters! 🎉🇺🇸🙏

MichaelDell's tweet image. Congratulations @jpmorgan on the opening of your new headquarters! 🎉🇺🇸🙏


Daniel Kaiser reposted

Everyone who signed the AI moratorium letter is now discredited. Looks like we have a bunch of new additions to that list.

A stunningly broad coalition has come out against Skynet: AI researchers, faith leaders, business pioneers, policymakers, NatSec folks and actors stand together, from Bannon & Beck to Hinton, Wozniak & Prince Harry. We stand together because we want a human future.…

tegmark's tweet image. A stunningly broad coalition has come out against Skynet: AI researchers, faith leaders, business pioneers, policymakers, NatSec folks and actors stand together, from Bannon & Beck to Hinton, Wozniak & Prince Harry.  We stand together because we want a human future.…


Daniel Kaiser reposted

People who don't know what they don't know think AGI is imminent.


The viral new "Definition of AGI" paper has fake citations which do not exist. And it specifically TELLS you to read them! Proof: different articles present at the specified journal/volume/page number, and their titles exist nowhere on any searchable repository.

m2saxon's tweet image. The viral new "Definition of AGI" paper has fake citations which do not exist.

And it specifically TELLS you to read them!

Proof: different articles present at the specified journal/volume/page number, and their titles exist nowhere on any searchable repository.
m2saxon's tweet image. The viral new "Definition of AGI" paper has fake citations which do not exist.

And it specifically TELLS you to read them!

Proof: different articles present at the specified journal/volume/page number, and their titles exist nowhere on any searchable repository.
m2saxon's tweet image. The viral new "Definition of AGI" paper has fake citations which do not exist.

And it specifically TELLS you to read them!

Proof: different articles present at the specified journal/volume/page number, and their titles exist nowhere on any searchable repository.
m2saxon's tweet image. The viral new "Definition of AGI" paper has fake citations which do not exist.

And it specifically TELLS you to read them!

Proof: different articles present at the specified journal/volume/page number, and their titles exist nowhere on any searchable repository.


Loading...

Something went wrong.


Something went wrong.