I typically prefer subbed; watching dubbed feels off somehow. Then again, I’m perfectly fine with the dubbing in DBZA so it must be a preference I learned somewhere!

If you want the FOSS community to be there for you, don’t go out of your way to use the corporate BS that’s actively destroying it.

Heheh, and another one of you thought to call me hyperbolic.

intentionally designed to cause disruption to a computer, server, client, or computer network

Actively and deliberately adding text with the explicit purpose of attempting to delete other peoples’ work fits this quite nicely, thank you for including it to illustrate my point.

Someone actively trying to sabotage their users is not okay, regardless of the excuses you want to trot out.

I’ll try to make my point clearer:

Damaging other people’s shit because you don’t like the way they live is not okay.

I don’t see how that’s relevant here. Banksy didn’t then go and shred every other piece of art the buyer owned. That would be pretty fucked up.

I appreciate you seeing this for what it is and saying so, where other people are twisting themselves into pretzels to pretend like it’s anything else.

Classy.

Yeah, I don’t care what their reasons are, they added malware to their own project to fuck over a portion of their user base. That’s not okay.

Let’s analogise: an artist discovers their art has been printed out and hung on the wall in the home of someone they don’t like. Does that artist have the right to try and burn down that person’s house in retribution?

Ahhhh fair play. I have a lot of freedom since I’m paying out of pocket for my own use. I have a pretty beefy rig for running local, but it’s not beefy enough to run deepseek pro and the like 😬 so, I have a bunch of subscriptions to try out a bunch of different models and see what works best in my workflow. I also have a problem with making alts in games, which seems like it rhymes 🤔

Been pretty impressed with glm5.1 too, before deepseek-v4 came out, but you’d be amazed what even a smaller older coding model can do with the right config and a little proactive context management. I really hope this trend of smaller, better models for local agentic use continues.

I’ve been using deepseek v4 flash on opencode’s infra for a couple of weeks and it’s pretty solid for something so low-cost. Honestly satisfied with it over Claude, for the premium Anthropic charges. Have you tried it at all?

Until we get to the total plan limits. They don’t cover costs to infinity, you know. And for something like cancer treatment, most plans don’t actually offer enough coverage to sustain it for years, so you’re still on the hook for tens or hundreds of thousands of dollars after all is said and done.

No, that guy holding a swastika flag at a Nazi march isn’t a Nazi. You don’t know anything else about him!

I don’t need to, though - and now, I don’t need to know anything else about the guy who said that, either.

I hate how transparent and low-effort your trolling is. Yes, there are Nazis in Ukraine. They’re not in power, and it still doesn’t justify Russia’s unprovoked war of aggression.

Now, before you spout off about literally anything else to avoid addressing that point: don’t bother. Instead, try harder. We’ll still see you for what you are, but at least it might be more of a challenge and you might experience some character growth along the way.

Personal ones. If I had to distill it here, I require that my prose be varied but sensible (e.g. breaks the rules for readability, never egregiously), with no formula, no waffle words, and no meaningless sentences. Tropes can be fine but only used sparingly, naturally, and absolutely not always in the same place. Reading level should be collegiate, but not doctoral, and industry terms should be defined on first use. These are stylistic questions shaped by years of reading professional authors and no AI has any idea how to do that out of the box. Reading what most people put out with AI, I can see the patterns like repeating textures, and it grates on my optic nerves.

I’m making progress on that with mine, but it’s a tough problem. Needs manual tweaking even after multiple rounds of automated revision.

I think you mean to ask

Do you seriously have standards for what you read?

And… yes, yes I do.

The sentiment is true - my employer is heavily adopting AI in our processes and we haven’t fired anyone, nor are we going to. We don’t have the numbers to lose, and everyone who’s using AI in their role wants to, because it’s improving our output.

But this article was written by AI and prompted by someone who didn’t try to improve its writing. Therefore I’m not interested in reading it. Why would I read a slop article from someone else, rather than having my own assistant compile a better one?

Why didn’t the person that died just not die? Are they stupid?

• A person who was not there and did nothing to help

The amyloid hypothesis has been supported by fraud for 20 years.

Pharma made drugs that dropped amyloid by 30% and they do NOTHING despite FDA approvals.

Okay, got any sources for that, or should we take you at your word?

Mouse models do not get dementia.

I don’t believe I said they did. Why the confrontational tone?

Misleading headline. What they actually demonstrated is reversing amyloid accumulation and the cognitive deficits in a transgenic mouse whose pathology is essentially just amyloid accumulation. Calling that “reversing Alzheimer’s” treats amyloid buildup and the disease as the same thing, which is exactly the conflation the amyloid hypothesis has been criticised for over the last decade.

Alzheimer’s in humans is amyloid + tau tangles + neuroinflammation + vascular dysfunction + actual neurodegeneration (entorhinal and hippocampal neurons dying, brain volume measurably dropping on MRI). Tau burden correlates with cognitive decline far better than amyloid does. The IBEC paper addresses one of those layers, the upstream-ish one, in a model that doesn’t reproduce most of the others. Fixing a cause in a young system before damage has accumulated is just not the same operation as fixing an established disease in an old human cortex that’s already lost the cells.

The human translation data backs this up. Lecanemab clears plaques and slows cognitive decline by about 27% over 18 months. Donanemab clears around 76% of plaques and slows decline by ~35% in early AD. In both trials both arms still declined, treatment just declined a bit more slowly. Northwestern’s Mesulam Institute puts it bluntly: “These medications do not reverse existing disease or stop the progression.” So removing amyloid in a system that already has the full human pathology bends the curve, it doesn’t undo anything.

What the IBEC team has here is a genuinely interesting result for the cerebrovascular angle, where BBB dysfunction and glymphatic clearance failure are upstream of plaque accumulation rather than a downstream consequence. The LRP1 transport mechanism and the multivalent ligand design are clever and well-grounded. The fair claim is “we improved amyloid clearance and rescued behavioural deficits in an amyloid-overexpressing mouse by targeting BBB transport.” That’s a real contribution. “Reversed Alzheimer’s” sells the mechanism by overstating what it did, and it sets up the same disappointment cycle the field has been through with every other anti-amyloid intervention that worked great in mice.

Original paper, for anyone wanting the actual data: https://doi.org/10.1038/s41392-025-02426-1