The more you try to stop them, the better their business gets, heh.

Perhaps it should be a new index that tracks a bunch of the major and minor indices

There are already whole-market indices they can invest in.

What about SiPearl and their big Rhea chips, Graphcore, and the big one , Imagination Technologies? What about the slowly withering Global Foundries?

I think Europe needs to pay attention to at their existing homegrown silicon gems and save/nuture/integrate them first.

But as with all things, popularity up, quality down.

I wouldn’t blame people entirely. Reddit’s engagement algos and UX design (aka rewarding short, biting posts/submissions) went a long way.

It says a LOT about where your faith actually lies if you have to promote a false reality to justify it.

The irony is that such fundamentalists rely on so much engineering, built on layers of scientific research, for what they do (like eating. And housing. And recruitment. And printing and distributing that textbook), and… yeah. It’d be like a flat-earther in orbit. It’s beyond ironic: it’s just not a possible situation without the help of outsiders refuting that belief.

I have a lot more respect for the Amish, isolated monks, folks that take their beliefs seriously and consistently in their lifestyle.

I would love a ‘star wars visions’ single season of one shot episodes like this. They could even vary the style every episode, which Bryke would love.

I suppose os.path is simpler? It’s a string and operation.

Python is all about ‘attention efficiency,’ which there’s something to be said for. People taking the path of least resistance (instead of eating time learning the more complex/OOP pathlib) to bang out their script where they just need to move a file or something makes sense. I’m with you here, but it makes sense.

…Also, os.path has much better Google SEO, heh.

In Tyler Boebert’s cases, Tyler was never the victim, but always a/the perpetrator, with his mother enabling him.

Where are you getting that from!? The tabloid article, which lists off crimes like a wreck, a messy theft ring involving drug use, and an apparent case of abuse from his own father, which have plausible but lenient punishments.

In January 2024, Tyler called the police and said his father was assaulting him at their home. Jayson Boebert was arrested after he allegedly “pushed Tyler to the ground and pushed his thumb into his mouth,” according to a Garfield County arrest affidavit.

This isn’t some giant crypto con or insider trading from the Trump kids, it sounds more like a struggling young adult to me.

What if Tyler’s mom wasn’t a politician, but, I dunno, a high level, highly paid engineer? Should his screw ups be under a microscope with “a moral imperative to consider every new crime of Tyler, to be news worthy” just because of what his mom does? And the possibility she used her weight to try and help him? Attention like that is a great way to screw up a struggling adult’s life.

I guess sending nudes is the norm and ‘expected?’ There’s social pressure to conform, and as we know, security isnt generally in the front of people’s minds…

No.

If it’s like nepotism or fraud related to their position of power, yeah.

But this sounds like domestic trouble not instigated by the politician themself. Bad shit happens in families; I don’t feel comfortable looking at that with a magnifying glass unless they want to share (or, again, it veers into abuse of power), and that includes checking if they’re leaning on police.

To put it in another perspective, let’s say you are a wholesome politician with a daughter in an abusive relationship. Her partner gets her into drugs. You think that should be plastered all over daily beast or breitbart?

What if you had a younger kid in trouble? Would you want the entire world breathing down their neck and publicizing every single interaction with law enforcement under a microscope?

I don’t think stuff like that should be public, even if the cost is missing some political pressure; politicians are human, that’s just an unreasonable expectation of transparency. And if that is the expectation, the only candidates you’ll get are liars and crooks used to skeletons in the closet.

This ^ is what I meant.

If you mean like a laptop, look for MoE models like Qwen3 A3B. And pay attention to sampling, try low or zero temperature first.

Well, since they’re open models, that’s easy to fix. And has been.

https://huggingface.co/perplexity-ai/r1-1776

https://huggingface.co/microsoft/MAI-DS-R1

Not to speak of finetunes that will do unspeakable things. It is not cost prohibitive or hard.

You want to decensor “Open”AI though? Tough. They don’t even offer completion endpoints anymore!

In other words, people shouldn’t and don’t have to use official Chinese releases if there is even the slightest concern of political censorship.

TBH I hate going after family like this, unless they’re explicitly inserting themselves in their parent’s politics. Feels gross.

The Deepseek distills will do pretty much anything you want in completion mode. Just modify their chat template.

My impression of Chinese models going back to Yi is that they’re more uncensored in English modes in particular, but much more hesitant answering in Chinese characters.

…That being said, Deepseek 14B is basically obsolete now. There are much better models in that size/speed class depending on your hardware, including explicitly uncensored ones.

Just to add to this, the biggest moat Nvidia has is not transistor density, but their software ecosystem.

Every since like the GTX 200 series in 2008, Nvidia stuff has been the standard for academic research, and it basically only works on their GPUs. Anything for research is done on Nvidia GPUs, which is tweaked for enterprise deployment on GPUs… if you want it on something else, you basically have to start from scratch. And dump a tremendous amount of brainpower put into optimization.

AMD’s in an interesting position here because they’ve been making Nvidia GPU competitors for literally decades. Their architectures are actually quite similar, hence it’s easier to ‘emulate’ Nvidia on AMD than pretty much anything else.

…That being said, the Chinese have made tremendous progress busting out of the Nvidia software ecosystem, hence these chips are actually being used for real work.

If they manage to actually get this into peoples hands

To be clear, I think they’re talking about mega-pricey server products, where the minimum size is usually 8 of them in a box. This isn’t a home lab kind of thing.

It’s not theoretical. They’ve already released an 300B LLM dubbed Pangu Pro, trained on Huawei NPUs:

https://huggingface.co/papers/2505.21411

And it’s open weights!

https://huggingface.co/IntervitensInc/pangu-pro-moe-model

It’s actually a really neat model: the experts are split into 8 ‘groups’ and routed so that the same number are active in each group at any given time. In other words, it’s specifically architected for 8X Huawei NPU servers, so that there’s no excessive cross-communication or idle time between them.

So yeah, even if it’s not a B200, proof’s in the puddin, and huge models are being trained and run on these things.