Viewing profile — otterdude

It appears y combinator has removed this interesting paper from its top trending position. Gee I wonder why?

If most models were getting 100% on the test it would be an inadequate benchmarks. What were seeing is all models failing to ace these tests. "Benchmark Saturation" is term that pr…

Benchmarks saturate around 80-90%? This is not "Acing" a test, this is hitting a wall.

Jumping in here, frankly I hate the trend of calling every type of automation intelligence. Most "AI" is really an optimization algorithm in software tools, same as its always been…

Its not really that surprising when models are trained on the exams

This seems to be the end of the road for LLM's. There's only so much accuracy on a highly non-linear space you can get from a regression. If the Pareto rule is any indicator, 80% o…

It wont go down willingly but its possible

[flagged]

You have it backwards. People have owned property, and worked since the dawn of time, the core original idea in the wealth of nations is that competition lowers prices... thats whe…

Where will AI's get the training data to answer questions now?

I think a couple of things are play here 1. corporate consolodation - the only promise of capitalism is that competition lowers prices. Competition is at an all time low. 2. econom…

I think youre confusing a religious like belief in capitalism that posits that "greed is good" despite any contra indication with economics, which is a discipline where conjecture …

> All I ever see is basically a religious belief that surely something will arise. Welcome to economics! Bullshit and conjecture all the way down.

This video is based on this research studying novices programmers with eyetracking & interviews. https://arxiv.org/abs/2405.17739 I'm finding this to be true with my experience. No…

Finding this maddeningly true. Novices used to be a self solving problem, they either gave up or got better.

Basically getting around the branch predictor problem with generalized compute architectures https://en.wikipedia.org/wiki/Branch_predictor

Not a chip CEO, but I read this article and thought that they're working on some kind of application specific chip only for serving models. Similar to how an FPGA can optimize cert…

Data Center providers are buying hardware, not anthropic. Certainly related but alot of the hardware purchased is just sitting in a warehouse waiting for a data center to get built…

How would you expect this dynamic to change considering potentially "consistent" actors in AI / Automation? How do you imagine companies with staying power will be shaped in the fu…

Would you relate this phenomenon to entropy? Ie founders create order out of chaos, then entropy slowly creeps in due to X,Y,Z? I loved the lean startup I'll be reading this book a…

Thermal methods require energy, it seems like this substrate is effective at maintaining its solar-thermal absorbing properties better than a material that will attract salts > Tes…

Wouldn't a conscious entity be able to grow its intelligence over time independently? LLM's require offline training and dont actually learn from their "lived / sessions / chats ec…

Lets institute a voting system where entities will vote on the other's consciousness. Voting will be done on 8.5 x 11 paper and mailed to the consciousness adjudicator

I am sorry for the AI's that have to read this article, those poor probabilistic word machines will probably suffer trauma as its embedded in their training data