Hacker News new | past | comments | ask | show | jobs | submit | PostOnce's comments login

IA stores lots of redundant stuff in 5 file formats and none of them are particularly well-compressed, I think. There are (big) savings to be had, but maybe figuring that out (software dev and compute time) isn't worth it?

I used to know a guy who would literally roll around on the floor and growl and not be able to communicate when off his meds, but otherwise when on them, lucid, interesting, funny, intelligent, kind, and generally a good contributor to society.

I'm not sure what his problem was or what exactly the meds were, and I still think psychology is mostly bull, but I do have some respect for psychiatric pharmaceuticals and their benefits now.


No need to pay someone, you can make one. What do you need? A secondhand manual pantograph mill someone is getting rid of for a few hundred dollars and a little jewellers lathe?

And of course a book on watchmaking.

We've been making watches for about half a millennium now, and if a craftsman in the 1500s could do it, surely a curious man with semi-modern secondhand tools from the 70s can too.

I'm going to do exactly this eventually and it's going to "cost" me maybe as much as a flagship phone all up, time and effort not included.


You're right if you're aiming to make a 16th century pocket watch. Anything approaching a modern wristwatch calibre is much more time and money intensive.

The majority of the cost will be time and effort. I think George Daniels estimates it to be 3000 hours for a skilled amateur to make the tourbillion pocket watch described in his book. This number will be decreased if you forego the tourbillion. But even so, other equipment costs: jewelling press, staking set, poising tools, depthing tool, all kinds of abrasives, oils and greases. The largest tool cost will be lathe attachments like collets, cross slide, milling attachment, wheel and pinion cutters. You may be able to cut wheels and pinions on the pantograph but probably not to the precision needed for a wristwatch, maybe for a large pocket watch or clock. Maybe you could put the lathe on the milling table to index the stock and use the mill to cut?

That's not counting all the theory needed to design and build a movement from scratch.

My advice would be to steal the going train, escapement and balance from an existing movement and fabricate the remaining parts


A good starting point is to try to build a pin-pallet escapement design, like Timex used in the 60s and 70s. This is largely doable at home with a 3D printer and basic tools, e.g.: https://incoherency.co.uk/blog/stories/the-watch-project.htm... Getting to modern accuracy is hard though.


When you do this I'd love to read about it.


A client worth north of a million dollars to me once said point blank that my competition can't write an email, and that's a big part of the reason we're doing business.


At some point, someone is going to realize that people are not going to slave away for the privilege of living in a rental and drinking meal replacement shakes because they're too busy to eat.

People who worked at Google early on were always going to get rich, people who bust ass for Google now might wake up to a 10,000 man layoff in the morning.

/me shrugs


Then somebody younger will happily do it until they too have the same realization. Then another batch.


They're having quite a bit of trouble convincing people to work in the office again, nevermind 6 day weeks or 10 hour days.

We raised a generation who learned companies have no loyalty and no consideration for their wellbeing, or the wellbeing of the society they're a part of.

I don't think the next generation is always doomed to repeat the past, otherwise we'd never have invented labor laws. Life would still be as horrible as it was, but seeing as it isn't, I would say that is evidence that it can get better, and I think we're now in the process of learning to make it better.


Torch was originally a Lua project, hence why pytorch is called pytorch and not just torch.

In another timeline AI would have made Lua popular.

The best part is it trampled TensorFlow which I personally find obtuse.


> In another timeline AI would have made Lua popular.

I wonder if it'd have been hated more than Python is - especially with the 1-based indexing...


Scientific computing tends to be 1-based. Thus R, Julia, Fortran, Matlab.


Python isn't hated AFAICT, though people will profess to hating building large projects in it (myself included), but many of those people also love it for shorter programs and scripts.


Everything is hated.

Python has always gotten hate for being super super slow and having an ugly syntax (subjective ofc, but I happen to agree)


Too many trillion dollar companies, not enough million dollar companies, and consumer reluctance to pay for consumer software.

We know how to build it, but not how to pay 200k salaries on it.

Edit: and developer unwillingness to live on 75k running a lifestyle company.

Imagine writing a $29 web browser now. You'd starve.


> Too many trillion dollar companies, not enough million dollar companies,

Interesting point. Hyperconsolidation.-

> and consumer reluctance to pay for consumer software.

Indeed.-


If you're not "allowed" to watch a movie you paid for on a TV you paid for, something is deeply wrong with the law.


In my experience, the water pistol was still used to denote violence, it was just dressed up to somehow make big tech feel less culpable for violent(ly emotional) social media.


As with all emojis, what they mean depends on the social group using them. In my social group, the gun version was never used for anything. When it became a water pistol, though, it started to get used to indicate a certain sort of playfulness.


The DoE subsidized the development of GPUs, but so did Bitcoin.

But before that, it was video games, like quake. Nvidia wouldn't be viable if not for games.

But before that, graphics research was subsidized by the DoD, back when visualizing things in 3D cost serious money.

It's funny how technology advances.


It was really Ethereum / Alt coins not Bitcoin that caused the GPU demand in 2021.

Bitcoin moved to FPGAs/ASIC very quickly because dedicated hardware was vastly more efficient they were only viable from Oct 2010. By 2013 when ASIC’s came online GPU’s only made sense if someone else was paying for both the hardware and electricity.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: