anjc

Brilliant video

If you were to post on another public website "My name is X, and my Reddit username is Y", would your anonymous Reddit posts now be considered personal data?

I suppose this would constitute indirect identification of a data subject under GDPR?

This is long solved

> I cannot understand how they think “being actice and doing things” is their selling point? It is just so unintuitive.

For me it's the best aspect of VR, a complete gamechanger. Maybe this is unintuitive from a non-user's point of view. The Vision Pro use-case of sitting on your couch consuming content is unappealing and off-putting to me.

> " By contrast, every demo that they showed was a person sitting on a couch by themself." > Wait, I thought an Oculus does the same thing too.

I never use my Quest sitting down. I use it as a roomscale device for, e.g., boxing, or playing table tennis or golf with family in other countries.

The Apple device looks great, but if it was the same price as a Quest 3 and I could only buy one, I'd choose the Quest 3, no questions.

> Even if that's the case, their paper should probably be more specific to the chemistry domain.

Why?

> The authors of the study are all from the chemistry department at the University of Kansas. Is this really the sort of paper they should be authoring?

What are you saying here?

Maybe I misunderstood but it seemed that he learned it was your fault, rather than this being an opinion he formed.

Are you going to raise this with the school? It seems highly inappropriate on their part

Brilliant and lovely, well done.

I think the applicability of techniques go both ways. Similar to what Sarah's dad said, I see research in ML now which rehashes expertise from many decades ago. Conversely, I see the demoscene implementing ideas on old hardware that, as you said, it would have been amazing if the technique was known at the time, but it simply wasn't conceivable then. It's only with advances in tech and lowering of technical barriers that new ideas are explored.

In any case, it's compelling to see people's expertise applied wherever.

Many comments are focussed on how this will affect models in deployment, but a more troubling implication for me is that evaluations and comparisons across models using past data are affected by this.

The article mentions something important and doesn't elaborate, that is, to trust the scientific method.

I've seen a few times state of the art developments that appeared to upend everything in my field. Only then with the passage of time do we then see that, yes the development is great for x, but older approaches are still better for y. With 'better' running a gamut of considerations, such as speed, accuracy, complexity, practicality, and so on.

ChatGPT is dazzling but I wouldn't be surprised if in 10 years we find that traditional NLP techniques have better precision for, e.g., language detection or information extraction than GPT-x (I'm making this example up), that evaluations from 2023 weren't rigorous enough, research was misleading, researchers were biased etc.

Hopefully cooler heads can prevail via the scientific method.

The approach you suggest is very common in the public sector, almost to a rule.

Public services are just massive, costly, complicated, and cater to all sorts of vulnerable populations, and the public have pay for it. The time horizon on some complicated projects is so long that knowledge can change in the meantime and make the project/strategies look incoherent. E.g., some countries decommissioned public transport options over decades when cars became more widespread, and are undoing it now because cars became too widespread, and it'll take decades to rectify. Then there are world events that upend everything, such as Ukraine or COVID-19. It's just very complicated, far more complicated than anything in software development in my experience.

Did you look at it? It summarises per-paragraph and the correctness can largely be verified by a layperson.

> the next time you have trouble booking a surgeon or even a gastroenterologist, you can remember that America’s supply of surgeons and gastroenterologists is being disproportionately used by the AARP crowd.

What a horrible sentiment.

Why are so many low magnitude stars in this image, not shrouded by gas, not visible at all in the previous image? Surely some visible light must be emitted by these?

> The lottery is an accepted tax on math morons, as my old teacher liked to say.

This glib old chestnut falls apart when you ask the "math savvy" what the probability of winning is when a person buys more than 1 ticket. Probabilities are unintuitive here (albeit simple) such that most people are "morons".

That's essentially what lottery/prize/premium bonds are, albeit they're optional

https://en.wikipedia.org/wiki/Lottery_bond

I presumed OP was talking about winders and so on. If shims were used for critical elements like optical alignment, I'm guessing that it's because this was the cheaper route to precision, rather than due to an inability to measure and manufacture precisely.

> Modern people overestimate how long we have been able to engineer things precisely

I think you're underestimating it also. For example, micrometres capable of measuring imperceptible differences have existed for many centuries, same for manufacturing machines (lathes, mills) that can shape objects to imperceptible requirements.

Cameras probably just don't require such precision.

It's state of the art. It sounds amazing compared to anything that's existed before. What alternative are you comparing this to to say that it's awful?

That's the perfect use-case for PHEV

If someone can't say "got a minute to chat?" or even "hello" to you without causing anxiety, you need a long break, a new job, or professional help.

Surprised at all the comments agreeing with this article.

Not technical writing, but the best book I've ever read on writing (including Elements of Style etc.) is "Getting Published in International Journals: Writing Strategies for European Social Scientists".

On the surface it's aimed at academics who are non-native speakers of English, but in my experience most people could do with reading it.

> I attribute the 'magic' to large scale statistical knowledge assimilation by the models - and reproduction to prompts which closely match the inputs' sentence embedding.

In effect this is how humans respond to prompts no? What's the difference between this and sentience?

People also fail to use logic when assimilating/regurgitating knowledge.

> It's 1T worth of... what?

US dollars

Great concept. I'm not doing so well though :) Seems like a good way to internalise notation

Grandmaster moves are likely more optimal than any guesses an average person will make

Can you show the distribution of reading speeds. If this font has a multimodal effect then the t-test mightn't be appropriate.

Same where I am. If you have a dSLR even a 15 second exposure will show you Andromeda with light pollution. Time consuming hobby though :)