taladar @ taladar @sh.itjust.works

Posts

1

Comments

2,384

Joined

2 yr. ago

Every couple of years a shiny new AI development emerges and people in tech go “This is it! The next computing paradigm is here! We’ll only use natural language going forward!”. But then nothing actually changes and we continue using computers the way we always have, until the debate resurfaces a few years later.

Reminds me a bit of graphical programming. Every couple of years someone comes up with the idea of replacing textual programming with some kind of graphical interface with arrows between nested boxed of various shapes and it inevitably fails.

Yes, or The Story of Mel

To be fair making outrageous claims with no relation to reality is sort of an all year thing with AI bros anyway.

Reminds me of that old magic/more magic switch story.

Oh yeah, completely forgot to list the destruction of aesthetically pleasing views with public advertising boards and the waste of lifetime spent on watching ads in my list.

Edit: also loss of life from depression and inferiority complexes caused by unrealistic life and body image goals in advertising

Wouldn't be so sure about the entire 21st century. There are some major efforts by the English speaking parts of the world in recent years to ruin that for themselves.

To be fair whoever invented YAML was just a sadist in general who wanted to see everyone in the industry suffer.

So only indentation, no alignment then?

On a long enough timeline, neither is Rust probably, but such is the price of innovation.

It is always so weird to me that people literally seem to believe that complex inventions like programming languages are something we got to perfection within 20 (in C's case) or 30 (in C++'s case) years of the advent of our industry. Especially considering an iteration cycle is somewhere in the decade or longer range for these. I would expect this to improve for at least a couple of hundred years before we reach the point where nothing new can be added to existing programming languages that is worth starting over with a new language to reap the benefits.

The Wikipedia page for SearX claims it has been discontinued in 2022. The GitHub repository being archived in 2023 would seem to support that.

You mean because my responses actually fit the comments I respond to unlike your low effort FUD that is phrased in such a general way that it could easily be spammed all over without attracting too much attention?

I don't know why you people all think AI has some magical capabilities. Tracking which aircraft is owned by whom isn't exactly a complex task if you have the raw data of flight movements (needed for ATC and otherwise independently tracked by plane spotters,...) and people movements.

So judging by how many terrorist attacks and similar events had perpetrators that were in theory known to the relevant government agencies that means those private vendors are now somewhere above the "incredibly bad at it" level?

You do not see the scope of the way advertising ruins our society.

We have

• a world-wide surveillance and profiling network collecting data about everyone for advertising purposes that is beyond the wildest dreams of authoritarian regimes
• ruining our planet's climate through advertising efforts by the fossil fuel and automotive industries among others
• ruining our health through advertising efforts by the tobacco, alcohol, sugar and similar industries
• destroying our systems of government through advertising efforts for political causes and for populist candidates
• keeping our population divided by using advertising methods to spread wedge issues
• destroying genuine grass-roots causes and spreading astroturfing to hinder popular movements, unions,...

Essentially everything you think of as propaganda is just applying advertising methods and societal structures for political purposes.

I guess it depends on the kind of job but at least with focus heavy tasks like programming I can already notice my error rate increase significantly by hour 7 or so. I can't imagine working 10 hours a day without spending half of the next day fixing stupid mistakes from being overly tired.

Realistically most people will still not run devices 24/7 at home. Data centers will always have a place for that kind of service, even if they host a lot of small, independently owned devices by the very same people whose home connections we are talking about.

I would say most people do not need a home connection that is in the same order of magnitude as the average data center server connection in use at the same time. Mostly because by definition there won't be many servers to transfer data from and to at that speed and the average person doesn't run too many connections in parallel.

Actually it is mostly the US that has been caught doing that. Not saying others don't do it but the US is the one who is bad enough at it to get caught regularly.

This isn't really a very high throughput anyway, some devices had half of that back in the days when 2-4TB was the largest capacity for NVMe devices so it hasn't really kept up with size very well.

Sure, six vendors all decided to release scam SSDs of similar sizes.