How does it make sense to say that speech developed 60 kya when Mitochondrial Eve and Y-Chromosomal Adam are commonly dated 200+ kya? Not to mention the recent advances within the past few years that more or less prove Neanderthals and Denisovans spoke as well. With that information Homo heidelbergensis almost certainly spoke as well and if I could bet on it, I would advocate that the smart money would be on Homo Erectus having some level of speech too. In fact a recent paper argues that case: https://link.springer.com/article/10.1007/s13752-025-00521-8 It's unfortunately not on SciHub though so I can't attest to it.

It sounds like people with terminal illnesses are getting thrown under the bus in order to justify banning open weight models, along with decreasing AI Capex in order to increase profit margins at American AI labs?

Is it just me or did this entire essay feel like it was penned by an LLM?

I really hope OpenAI doesn't take the bad press some people are giving them too seriously here. They should throw their whole weight behind the rest of the Millennium Prize Problems. To think – if everyone lets their egos calm down we could have the Riemann Hypothesis solved by the end of the year...

Super nice to be able to finally have voice mode search stuff in the background. The more conservational vibe is also a nice touch. The real game-changer we're all waiting on, though, is being able to vibecode/deeply interact with an OS through voice mode. Until then voice mode will continue to be a niche, interesting product. Once that's implemented, though, it will be one of the most important technologies available. There's a lot of people I know that don't use a lot of the frontier capabilities of AI right now because of the required typing interface. Once that's changed and you can interact with these capabilities through voice mode (and ideally on your phone) I definitely see them and lots of people like them becoming big users of AI.

athyuttamre is a coward if he doesn't answer this

The interesting thing about models this small is they should be able to be put on a single Taalas chip (the HC1 already runs a Llama 3.1 8B model). We're already at the point where half-decent reasoning could be run on an ASIC (and at mind-boggling speeds).

Yeah, if they can fit an 8B model that's really good at improving the output by thinking, running at 16K tok/s on Taalas would be mind-blowing.

Given this and the quality of open models, it makes no sense to me that there’s a future for Anthropic et all?

Packaging a capability into a consumable form will still be business.

It's like web hosting; all the open source tools are there and free, and yet website tools, hosts, etc flourish.

It’s true, but hosting prices are still within spitting distance of rolling it yourself.

SOTA providers are expecting some level of margin. Companies everywhere have a tight eye on their AI bills right now.

The motivation is there if the models get good enough, even if it’s more painful.

Exactly, it's like web hosting.

And I don't see a lot of web hosts with twelve-digit valuations.

it's not like anthropic will stop improving their models

If the way someone is rejecting a trans person is considered a hate crime, maybe that someone is the problem? If someone were to beat you up or misgender you just because you're Southeast Asian that would kind of also be a problem.

Exactly. It's so interesting seeing the unstated assumptions software engineers let slip out. It really reveals the bubble they live in.

PurdueX has had some really interesting courses. I wouldn't recommend paying for them though since you can't deduct them like normal tuition expenses.