alienbaby

I guess then it has no problem working in high dimensional spaces, why would it have?. I wonder if perhaps it's pulling on a bunch of those dimensional threads and getting the outcome it's after, while subtly altering the outcomes across many thousands more dimensions in unnoticed ways. This feels more like an us problem, as in, we think in 4D (just to pick something) and we imagine and theorise in 4d, and propose questions in 4d, whereas a model has no such barriers, and can think in many more D's than we can.

This has to be a good step to recursive self improvement, no? Give a model it's own weights and tell it to make them better at x,y,z, etc...

nobody is saying anthropic was hacked?

Ok, but how about student grades? I wouldn't have expected any impact on mental health resulting from a school ban considering the exposure they get every other hour of the day.

why not just say 'millions'

"Our next iteration of this scheduler aims to scale to millions of GPUs"

o7

Blades of glory for the original quote.

as has happened time and again through history. software development skills will fade as models become genuinely better at creating, designing, architecting, engineering, testing, validating code at the level of fully complete products.

There will still be the need for humans in the loop to make sure the final product is suitable and what the humans genuinely wanted. We'll all become QA / Acceptance testers. For a little while , that will go away in the end too I think. software will become nothing more special than every day household objects.

isnt that kinda like downloading and using software a team you have no knowledge of has produced and released?

All you can go by is 'does it do what it says it does'

I think he means, if an average tree adds say ~10g of wood per day, I had 20B trees and cut down 250000 of them, how long do the remaining trees take to grow back the wood that was lost from the trees that were cut down.

Turns out, if the trees lost are about 2000kg, it takes about ~ 2.5 days.

am i missing something or have they updated the page? the blue highlights and grey arent that bad to read?

The article implies they are wrapping up their current models, and instead aiming to go all in on a frontier class model?

"they have gobs of money, "

They do? They are a spec in the ocean relative to big players offering inference and training their own models. Pretty sure one thing they are struggling with is the money to compete successfully in the ai game this time around.

That's what I've been doing in areas I lack knowledge tbh. Instead of a half assed attempt at describing what I need, I ask an llm to take my base idea and write it up using industry standard jargon as if I were commissioning a team to do the work. Has been working pretty well so far!

ever since I heard this as a young teen for the first time, it never fails to make me grin

'unlock prompt' .. err, ok. straight to signup. Smells like I'll need to pay for it, and I'm just curious to see one prompt from the front page. aand I'm done.

oh I wonder if I can point claude at it...

Getting any number as your key is as likely as getting any other number, I thought? 4 is as likely as 64257383994662784857722891857 etc ..?

Not just technically possible, as likely as getting any other random 116(?) digit number

have you looked closely at the state of your institutions lately. 'The country keeps running' .. a large thing has quite a lot of momentum, it will travel quite far from that alone even if you do remove it's kneecaps.

unless you dropped the '/s' ?

from emory (3 days.. stretching it ;P) the first block of links taking you to PEP's were broken, and I spot checked 'other links' down the page.

As I understand it, AI 'neurons' have nothing to do with IRL neurons other than a passing nod to the name loosely based on the idea that 'they are connected to each other'. A neural net is nothing other than an (very) approximate model loosely based on our understanding of IRL neurons. There are many many ways a real neuron could be resonding to it's environment to do what it does, whereas our artificial 'neurons' are just a mathematical function.

I would image then, trying to imbue our neural networks with the properties of real neurons is a futile ambition?

I assume it does get stuck, and creatures go extinct as their environment changes around them and evolutionary mechanisms don't by chance produce a change that enables survival either quickly enough, or because the required changes are not possible from the current state of the organism.

"If we subsequently discover"

hmm. What if they didnt realise at first because the code is actually fine and works as intended. They gonna back it out just because it was AI?

30 macs in a Cray looking case?

The 'look awesome project with funky name, here's how to install off you go' without a word to what one earth it actually does is amazingly common. Plenty of things posted here I bounce away from after hitting the linked page, finding something someone is obviously very proud of, but not having a clue what it's actually about :p

I've wondered how; "The Navier Stokes results used millions of dollars of tokens and thousands of parallel agents to get the result." can possibly not also include many dollars worth of duplicated effort? I find it hard to believe every agent was doing something novel. What this means beyond wasting dollars I'm not sure of, but just flinging around the numbers I don't think should be considered impressive or even required to get the results found.

A large number of the links in the page appear broken for me :/

is this what other peoples meetings are like? huh.

now tell me it's all AI too :D I can't tell, LGTM

The code documents the code, not the a hardware and software infrastructure within which you are building the code.