nbardy

I think it's weirdly just a choice of deciding to cut releases.

We already know OpenAI has "bel" that is MUCH better than astra and is being used internally

Have you even tried the new models? Opus 5.5 is a clear leap. Go back a year ago and try them and tell me there is any sort of plateau.

I’m pretty sure 99% of what people perceive is the old model training on the usage logs wherever they were stuck.

Step 1. Model can’t do something challenging Step 2. You try a bunch and fail Step 3. Anthropic trains on your usage data. Your current code base and current problem are now in domain Step 4. Model comes out and you’re shocked when it can tackle the thing you were stuck on Step 4. Codebase drifts significantly and you try new problems you thought were a similar level. Your code is less familiar and the problem doesn’t have a bunch of failure cases in the train set. Feels of it being worse on similar problems

You can estimate the model size by looking at tokens per second and comparing to open source models

One of the amazing things is that when every has one GPU, they will actually have 1k-10k agents at their disposal.

LLMs and KV caches have amazing performance characteristics with concurrent throughput. It scales very non linearly. So the token throughput within a batch scales WAY faster than the tokens per second of each user.

This is the reason the LLM providers have such crazy margins on their costs.

They had incredible revenue growth the last few years and just broke 100M in revenue. I don't know what their internal spend was , but that was almost half of their recent round in ARR. Mostly likely they were profitable or on a clear trajectory to revenue growth. Huggingface hosts a lot of data and models, but mostly static cold storage is pretty cheap tbh.

This is a wildly incorrect and myopic view on the world.

Finetuning model is cheap and incredibly useful for deployment. You don't need to pre-train a frontier llm from scratch to make useful models.

There is tons of domains where you and fine-tune llms and deploy them for value in companies and for your own entrepreneurship ambitions. I have made this a big part of my career for the last few years and now I'm working on finetuning models for starting my own companies.

The real revolution is both. The cost and capability of frontier intelligence will go up AND the cost of "good enough" intelligence will go down.

I think this is largely true. Theoretically it follows that mature industries would become this way more and more.

In my software engineering jobs there was always strong top down product direction. In my AI research jobs people are often staring at me blankly waiting for me to tell them what we can do.

A young industry the only people who really understand the potential of what the technology can do are the experts and practioners.

Overtime all the best ideas get picked off, and the general population who are non practitioners gain enough knowledge that they can understand what the technology can do better than the experts how can impliment it.

I think a lot of it is just time. The quality of a model is E * C

Where: E = Efficiency, and efficiency gains come from quality of data, quality of algorithms. C = Compute (Size of model, flops of train run)

So a better company can train a bigger and better model with less required compute which let's anthropic get there first. If another company does the same thing with a worse: model architecture, kernel, optimizer, etc... They will get there as well if they just run there train run with more flops for longer

Mythos was actually ready about 6 months ago. So if you have 6 months later or hardware setup and time to train you can get a lot done.

Everyone keeps repeating this who doesn’t understand the underlying technology.

Small llms are still way more efficiently server on big GPUs.

Sharing server capacity takes advantage of the massive parallel throughput and sharing of memory bandwidth.

You are sharing the GPUs with thousands of concurrent users.

he got promoted to chief scientist and them being an expert at all modeling will really help compared to being llm only

yea, a lot of my prior work was in the "code is the easy part" I was a frontend engineer for years. And something like 90% of my job the code was not the hard part.

I loved writing GPU shaders or optimizing visualization performance, but most of the time it was wiring up netcode to UI elements that exist.

Ironically as I've moved into focusing on more GPU and kernel programming AI is now lapping me there anyway, however the impact of knowing what sort of algorithsm are state of the art in papers, what is causing memory bandwidth issues etc... does a lot to drive the machine.

The faces look amazing. They can post train for identity preservation easy, that is just an adapter

Then they’re legitimately getting better at svg which is a valuable tool.