dcchambers

There's no doubt big tech is/was bloated. It was (relatively) cheap to stack and retain talent just so your competition couldn't get them. There was also an era where VPs and directors had direct incentives to grow their orgs as large as possible.

The opportunity cost is higher now (companies are actually spending money on capex), and changes in productivity due to AI mean that simply hoarding talented developers isn't as useful any more. Before, you could keep people "busy" doing things that didn't really produce anything useful, and still find ways to justify it. VPs and directors are now being rewarded for "doing more with less" - not just shipping the largest organization.

I think overall SWE numbers will continue to slowly climb, but I think we're going to see more big tech unwind bloated staff numbers.

Wow.

Looks better, thanks. What's the source for this data? Might be nice to declare that somewhere on the page as well.

My eyes have no idea what to focus on.

It would be more readable if every AI-generated page didn't go for the dark-mode + glowing neon cyber aesthetic.

My 2¢:

- It's been proven now that the "frontier labs" really have no technical moat. - DeepSeek has shown how cheaply inference can be done. - This creates a world where "intelligence" is a commodity and pricing is a race to the bottom.

This is not good for the multi-trillion-dollar hopes of the AI labs whose valuationa depend on them eventually reaching "ASI" or "AGI" or "RSI" or whatever they're currently clamoring about.

I personally believe this is a good thing. I wish it was coming from a US-based company, but DeepSeek is so far proving to be the OpenAI we all wanted.

It's helpful that DeepSeek is so cheap that it's actually reasonable to pay standard API rates.

Dax has said time and again they don't quantize the models they offer. Assuming they trust their upstream providers (I don't think believe they run their own inference) and we trust them, what you get with OpenCode Go is exactly what you'd get from the DeepSeek API.

I am personally a very happy OpenCode Go user. $10 month has never felt more useful.

At this point I am way more excited by DeepSeek releases than anything from OpenAI or Anthropic.

The speed and cost/intelligence ratio is just outrageous.

I don't think this is going to change anything for most people, but it certainly fits a niche for US-based companies that refuse to run Chinese models and only want to run their own inference.

But I'm really happy to see Meta do this and I hope they continue to release open models in the future. Open models are really the only way to ensure fair access to AI for everyone.

This is exactly the type of niche service that modern technology enables. Without that tech (3D printing, the internet, AI) things like this are simply more likely to just not exist because it would be impossible to make the economics work.

This is a net gain all around.

At this point if people keep buying Apple products expecting Siri to get better, that's on them.

Their official documentation says it's available to Pro subscribers in the Gemini app only, and available to Ultra subscribers in Antigravity.

Not available for AI Pro subscribers in Antigravity. What a disappointment.

Google...why. You keep getting so close but constantly fumble at the last moment. I can use Opus 5.5 and Sol/Astra on the $20 Anthropic and OpenAI plans.

You are already playing from behind. You need to GET YOUR MODELS INTO MORE PEOPLES HANDS by any means possible. Continuing to make it difficult to use your products DOES NOT HELP.

This model might be great but the majority aren't going to subscribe for $100/month just to try it.

> Meta could easily show how far ahead its own model is against the rest of the pack of open models but they chose not to. Why?

Wonder is not an open model and there's no claims that it will be in their announcement.

Despite the claims that they would release weights for Muse Spark, they haven't done that yet either.

> Luna performance

Source? On the available benchmarks it looks like Sol 6.1 level.

Super interesting that they're not even making Wonder Full available outside of their own coding harness, and thus not providing an API price to use in comparisons.

> these AI companies need to start thinking outside of these basic chat UIs, they do some things here and there, but its really depressing how little they do to innovate in these spaces. Same with the coding harnesses, the UI for all these things could be drastically superior.

I would prefer if the AI companies stuck to just creating better models and making them as cheap and accessible as possible. Let others build the products. I don't want 1-2 companies to own every product in the world.

Begging Anthropic to let us use Claude subs with harnesses other than Claude Code at this point.

This is awesome! When I was a student at UW I worked with the CHTC (Center for High Throughput Computing) and we provided a ton of compute to the Ice Cube project. Happy to see something big come from that!

https://htcondor.org/featured-users/2026-10-06-francis-halze...

Congrats to Francis and the whole team!

Of course it's not even available yet. Google - with all due respect - how in the world have you not figured this out yet?

We will soon look back at these companies like we did the pharma companies that created the opioid crisis in America.

And by that I mean almost no one in charge will be held responsible.

GPT-6 Sol released a week ago. Shortest model life ever?

Happy for Adam and his team.

This is so frustrating. Fandom is a cancer on the web. Independent wikis are treasures. Google needs to fix this.

I don't love this. Wish Valve itself would have supported SteamDB.

> "loyalty"? weird word choice.

All I mean is that none of the frontier LLMs are significantly better than the others for the vast majority of work, so devs are free to move between providers freely.

It's not the singularity.

One of the major LLM providers goes down for some reason. Traffic shifts to the other providers because devs have no loyalty. LLMS are commodities. The other providers can't handle the increase in traffic and they go down as well.

Can you use Pi with a Google Pro AI sub or do you need to use API billing?

I would really love to be able to use these Gemini models in Opencode or Pi with my existing Google AI Pro subscription.

Given the choice between a walled garden like Apple's app store or this, give me something open and more risky any day of the week.

It's easy enough for Omarchy folks to add "verified" or "trusted" developers down the line.