sottol

My guess is that established companies (esp. those with existing offices) will require on-site work, new companies without will more likely not... more flexibility for founders and reduced costs if they believe in remote.

That should leave everyone with plenty of options, but on-site + startup combo would have fewer positions and vice-versa remote + established company.

Otoh, over time I could see a profound shift to remote if these new companies were to start replacing old ones and remote worked for them. But there might be a limit where a certain company size favors on-site.

Agree, thus the scare-quotes.

Maybe a fire safety violation? I think windows can count as escape routes and probably need to be open able quickly.

I think it's just an attempt to stop "grey-market" imports if I had to guess and this old lens just got swept up in the dragnet. They're probably (ab-?)using ebay's IP protection systems though.

Manufacturers sometimes have vastly different pricing based on region and want to protect profits, to the point that most (all?) manufacturers will not service "gray-market imported" lenses in the US. So large 3rd-party grey market importers like "E-Infinity" even offer their own warranty.

The "win matrix" (dissimilarity Matrix) seems very interesting, looks eg like Vicuna13b paired against gpt4 wins 20% of the time. Larger difference than I'd have guessed based on scores.

Imo that would leave Italy and Spain and their respective union members in the dust competing against the US, China and India.

But maybe that's worth it, or maybe the Brexit will turn out great like pg said due to them not having stringent EU-type regulations on things like AI and banking (though if you're a founder leaving the EU i don't see why you'd go to the UK over the US or Canada).

What's the alternative? Brexit for everyone?

It does sound good on the surface. Equality for all. Right? Only selectively. But nice touch invoking immorality.

Is it moral to fly around the US and the world while the guy that's hiring the chef that cooks your food stays grounded?

This feels like the epitome of capitalist short-sightedness. There must be a way to build loyalty with your pilots by treating them well during the downturns and spending more on training pilots? It feels like there are massive layoffs and cost and pay reductions every 3 years, just to be followed by a boom shortly after and complete shortage of pilots.

Drum brakes have been around for a century, afaik they're more robust than discs but "lower performance". These look like a variation on a pretty standard design at first scan.

So I was wondering why they'd openly admit in a WSJ article that they've "watered down" their furniture and are doubling down on their reputation for flimsy particle-board and paper construction.

Also, I've observed this trend starting way earlier than the pandemic and the start of recent inflation, I bought their butcher block countertops in 2013 and 2018 for my workshop. Let me just say I ripped out my 2013 ones and transferred as much of the old ones to the new workshop as I could.

I can see two reasons for this article:

1. Their customers are noticing and it shows in survey and satisfaction and this is an attempt at curtailing the reputation damage (but don't really see why it's in the WSJ then, maybe it'll show in the coming quarterly/yearly's and they can point to this later?),

2. or they're signaling to investors that they're going to boost profitability and/or margins.

Either way, not really a positive trend in my eyes.

But also unfortunately, it's not like the alternatives are generally much better :/ Seems to be an industry-wide trend.

EDIT: nevermind, Ikea is entirely privately held -- no investors to placate, so 1. maybe?

Classic attention is quadratic in context length and faster alternatives seem to not perform as well, wonder how Hyena is in comparison to linear attention algorithms.

Not what I'm seeing from the linked map -- I see a 179 in the US and doubt the average is >180 in Europe? Eye-balling it, Europe looks about comparable to the East-Coast numbers, maybe +10.

Also, afaict that map shows recent, "live", up-to-date numbers, I assume seasons and day/night play a big role there.

But what are you mostly "teaching" the LLM then? Mundane everyday stuff? I guess that would make them better at "being average human" but is that what we want? It already seems that prompting the LLM to be above-average ("pretend to be an expert") improves performance.

Something about this also makes me feel a bit, hmm, I guess "concerned". I know SpaceX is very efficient and that they hustle to a degree, but their launch site looks very "min-viable" (at best). I hope this does not and will not extend to other parts of their designs.

In fact, it's probably a lot worse. EVs need better aero to extract as much range as possible out of the batteries due to the lower energy density.

Everyone? Imo more than half are worried, some members are thinking about switching careers.

Two random examples by searching algolia for a minute (entire posts, not just single comments)

https://news.ycombinator.com/item?id=34008901

https://news.ycombinator.com/item?id=33884202

Did they release the merged weights, yet? I'd love to try this model.

Afaict from the docs, you still need to request the original Llama weights from Meta (or get ahold of them another way), then apply the diff-weights requiring 60GB RAM?

Not as far as I can tell? Any examples?

Also, YC seed isn't gonna get you very far when compute alone for training an LLM is in the millions currently.

> I'd be surprised if Apple hasn't allocated at least $100,000,000,000 to AI development at a minimum and likely similar numbers from Amazon, Google and Microsoft

> Its not hype.

> ... AI to pretend to be a Furby...

Sorry to pick on this and be a little snarky, but all of this is hype.

100B? That's more than will ever be spent on fusion research - unlimited cheap energy.

OpenAI, the leader in the space, is valued at 29B and they've delivered a compelling product with seemingly less than 1B raised.

I get that people are excited, but this will be imo a smartphone type innovation. Sure it's made a lot of stuff sort of different (but not that much imo), lot's of new applications and apps - the "world's entire knowledge at your fingertips" but has it really changed humanity? Most people play freemium games on them, order takeout or take selfies and browse social media.

So I'm guessing this new wave of generative AI will mostly give us more of our shallow entertainment... virtual celebrities and gossip, on-demand personalized games and chat partners.

> None of those are perfect, but it’s better than Creality selling Prusa clones in every Micro Center.

Prusa i3 themselves are evolutions of the reprap Mendel design, so I'm not sure if it's fair calling creality enders prusa clones.

Vorons are pretty terrible at PLA, not enough cooling and enclosure. They're great Abs printers though and PC blend.

As far as I've seend the X1s lidar is mostly useless, it's doing fine-calibration at most based on the current filament profile and seems to be inconsistent.

That's not to say that the X1 itself is - just that lidar is imho a gimick.

> Surely you have to do some kind of math eventually, or else you'll end up prioritizing whatever sounds the most dramatic instead of the things that actually matter (consider people who are more afraid of sharks than drowning, even though shark deaths are 1 in 4,332,817 and drowning deaths are 1 in 1,134)

TSA? Or any other "but think of the children"-type bill? Or look at water-scarcity "solutions" in the south west for something different.

I'm serious, "prioritizing whatever sounds the most dramatic instead of the things that actually matter" is literally 90% of the politician's playbook.

Taking the title literally... that's more than 5% of the world-wide working age population -- I'd be surprised if that wouldn't cause luddite riots. But there are a lot of weasel words in the article and even the headline has a "could", I guess the gist is that they really expect generative AI to lower wages. Wish I could my hands on the actual report and not just this article which seems to summarize the reports abstract at best.

Thanks, I love hard numbers.

The world bank currently pegs working age population 15-64 at 5.12B, so that'd mean chatgpt would put nearly 6% of the working age population out of a job (ofc depending on your definition of "displace").

Blind-sided how?

50% of software engineers unemployed in 2 years, 5 years, 10 years?

Or something like 50% of the US/world population using a chat-gpt style assistant to perform a complicated mental task (aka not "Alexa, set a 2 minute timer.") once a day/week in 2 years, 5, 10?

Don't get me wrong, chatGPT's emergent intelligence is impressive and I'm playing with Alpaca 13B and other models locally, but I'm not sure it's going to be as transformative and as quickly as many here seem to think. Humans and society are inherently resistant to and slow to change.

> Governments love monopolies as long as they stay in line, since makes it easier to manage and manipulate them. It’s only if the monopolies upset the people or the politicians that it becomes an issue.

Very true. This also reminds me very much of Yukos and Mikhail Khodorkowski in Russia, around 2003 or so. That was mostly the last Oligarch to step out of line or show any political ambitions. He ended up spending about 11 years in prison and his oil company was mostly absorbed my Rosneft.

Then I wouldn't be surprised if they also had a long-term, iron-clad license at fixed cost.

Both Google and Apple have multi-arch support and Apple has pulled off a ISA transitions, guess that really only leaves Apple's investment in the actual M1/M2 chip-design itself.

Apple's price-point and margins are high enough but they're pretty protective of their margins afaik. I wonder if it's worth transitioning away though.

I think this will definitely boost Risc-V investment and maybe other alternatives, if for no other reason than to have a viable threat to leave ARM if they decide to get more aggressive in pricing.