Real gamble, that. You're injecting something that no regulatory body that said was safe and that that supplier is giving you what they say they are? Just like injecting whatever it is they happen to give you?
If you're doing this sort of thing, usually you batch your orders together (ideally with some friends) and randomly lab test a few vials selected from the large batch you receive. The powder form vials you receive before mixing will last more or less indefinitely in a freezer, so you can order multiple years worth of supply at a time.
The potency and purity are typically pretty good but from time to time vendors will be outed for either outright fraud or a bad batch that is underfilled or low purity. Most of the direct vendors (out of China) will guarantee fill and purity, and also give credit on your next order for lab test costs if you share the tests with them for future marketing purposes.
All sorts of risks involved of course, but you can mitigate them fairly effectively. Your risk tolerance may vary. Some folks will buy single vials from US resellers and trust those seller's lab test reports. Others (smartly, imo) don't trust that and do their own testing directly on the product they receive.
This grey market supply chain is remarkably robust these days. Plenty of suburban Facebook soccer moms organizing group buys.
No, you can know that decisions are stupid before knowing their outcome. In fact, decisions can be stupid even when they result in a neutral or positive outcome.
the worst case for this is if that doesnt cause sufficient problems to force a reckoning. it has to hurt badly enough that we put significant effort into training legibility into the process.
the equilibrium state would be having found a good pattern for making patterns good. we probably aren't too far from that, so i think your worry may show up in the short term but not the long one
that would do it too, and i suppose people are working on that. but even without that, there arent infinite low hanging fruit problems to find, and any given code can just sort of be play-doh'd into other code. so the llms will need a few idk years worth of rounds of being trained on the things that pop up that wed never have the time to harden, and things will settle. like, its hard patching stuff up quick right now but we're managing and society hasnt fallen in to chaos.
What I have been reading makes me want to both agree and disagree with your take here. I think the fact that some of these things are not only established but are much simpler than one would think at first glance yields the much cooler realization that you can not only do a lot more than abliteration, but that humans can do it by hand if the thing is made legible.
From my experience with models when it comes to discernment about particular things, the models, even the best ones, still are not very good. The thing that we are most missing I think is acceleration in creating editing tools for humans that make the otherwise inscrutable both visually legible and interactive.
At least the Qwen3.5-0.8B loop bug isn’t that special (or rather not something worth extrapolating). I literally had this issue last week and told Astra to just go fix it.
Thats not what the issue is, it's that they aren't free to job-hop in the way that other people are. H1B keeps their wages artificially depressed and they have a sword of damocles over their heads threatening deportation.
It's an incredibly obvious optimization on the autoregressive model, and converting a training pipeline to adopt it, with how good autoregressive models are now, is as easy as saying "do this but jepa" to codex or claude. It's gonna be like how everything is becoming rust now.
Oops, HybridLink as a paper came out little bit ago, and it was an architecture I more or less converged on in my own work. JEPA is the obvious successor to autoregression once you look at it a little bit - to the extent that you can take a non-jepa transformer and just slap a couple things on it to make it a latent reasoner without crossing the token boundary. It's faster and more accurate, as you would expect from not having to project down to tokens and have the model reconstruct its own latents.
Man, the "protests don't work" energy is way stronger here than I would have expected! Whatever happened to the spirit of disruption or going a different way?
reply