Theory 1: Some increasingly-large split of inference compute is moving over to serving the new model for internal users (or partners that are trialing the next models). This results in less compute but the same increasing demand for the previous model. Providers may respond by using quantizations or distillations, compressing k/v store, tweaking parameters, and/or changing system prompts to try to use fewer tokens.

Theory 2: Internal evals are obviously done using full strength models with internally-optimized system prompts. When models are shipped into production the system prompt will inherently need changes. Each time a problematic issue rises to the attention of the team, there is a solid chance it results in a new sentence or two added to the system prompt. These grow over time as bad shit happens with the model in the real world. But it doesn't even need to be a harmful case or bad bugged behavior of the model, even newer models with enhanced capabilities (e.g. mythos) may get protected against in prompts used in agent harnesses (CC) or as system prompts, resulting in a more and more complex system prompt. This has something like "cognitive burden" for the model, which diverges further and further from the eval.

alias yt-pl='yt-dlp -o "%(channel)s/%(playlist_title)s/%(title)s.%(ext)s" -a playlists.txt'

I am a software engineer, I also runs a small family farm. I have 3d printers and laser cutters and lots of aluminum extrusion and raspberry pis... but I keep those things indoors, away from the dirt, sun, and rain. I can't imagine a real farmer using a contraption like this. Tools have to be reliable to last. I have to replace my solid steel shovels every few years because they wear out, how is this supposed to work?

How many of these passing tests are still actually testing anything? Do they test the tests?

So, if they IPO, will they first be distributing equity to the community? Or are they just cash-grabbing a retirement play?

Can anyone recommend one of the foss gallery apps available on f-droid?

These ANC headphones didn't support Bluetooth at all, while other models with the same name did. In 2024, for $100, a wireless headset that only functions with the included dongle?! I made peace with it and got used to swapping the dongle to each device I use. Because of the poor dongle design it couldn't fit in my phone without removing the phone's case, so I listened to podcasts with my case off. Within the first two days the dongle just completely stopped working. Complete garbage. Not only that but the audio quality was horrible, the microphone sounded like a rotary phone, they were huge and bulky and hot to wear. When I contacted support asking for a new dongle they emailed me back the same template six times from three different representatives as if I hadn't answered all of their questions each time, and they wanted me to send the entire headset back to get a replacement.

I left the experience thinking "Wow, this garbage company must hate their customers." Even after all that, I didn't realize they hated their customers enough to install bitcoin miners on their mouse drivers or to sell them fake N95 masks. The punishment here doesn't even come close to justice, if anyone got sick the executives in charge of bringing those masks to market should be held criminally liable.

The other thing is that words have a lot of power in the cultural frame, even just the concept of aloha being something that could be "unleashed" is likely to offend.

All to say nothing off the palpable fear people have here of robots taking hospitality industry jobs like housekeeping (which are unionized in many hotels out here, and are actually one of the few low-barrier-to-entry jobs out here that can support a reasonable quality of life)

I'm sure I'll get a ton of downvotes for bringing up cultural sensitivity and pointing out these concerns -- I don't mean to imply they're all 100% rational nor that no one should say "aloha" unless they're Hawaiian, but if anyone at DeepMind had a Hawaiian cultural frame I think they likely would have flagged these concerns and recommended a different name.