this post was submitted on 27 Dec 2024
306 points (95.5% liked)

Technology

60332 readers
4450 users here now

This is a most excellent place for technology news and articles.


Our Rules


  1. Follow the lemmy.world rules.
  2. Only tech related content.
  3. Be excellent to each another!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, to ask if your bot can be added please contact us.
  9. Check for duplicates before posting, duplicates may be removed

Approved Bots


founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
[–] [email protected] 1 points 1 week ago* (last edited 1 week ago) (1 children)

And if you don’t always double check, it will bite you in the ass eventually. Good luck with that.

When did the web ever present itself as a completely factual and never wrong? There's plenty of evidence of wikipedia being wrong on wikipedia

https://en.wikipedia.org/wiki/Wikipedia:List_of_hoaxes_on_Wikipedia

Do I get things wrong? Sure, never said I was perfect either, if someone tells me I got a stat or a figure or something wrong, great!

The question for me is: is it wrong enough to make the results completely unreliable, and the answer to that is no, more often than not it provides accurate information.

If it wasn’t for the web being a monetized SEO algo shithole we could still just search the web!

That's not accurate to me, AI/SEO search results are still a minority of results that I get, most of the time I get close to what I'm looking for, but AI search summarisation is essentially the next level of search for me:

Dogpile/Altavista/AskJeeves > Google > AI powered search summarisation

I get essentially what I'm looking for directly, why click on a page with 47 ads, a video pop up or something else when all I'm looking for is:

https://www.perplexity.ai/search/do-you-have-a-basic-egg-on-toa-pkpsq9WwSMm5G8ICsmDnbw#0

Is it a complete replacement? Not yet, Ecosia is still my daily driver having used it 25,000+ times in the last year but AI is making a serious dent in how often I use it.

we don’t need to build acres of compute powered by nuclear reactors to fix the problem.

I would keep an eye on that, the gains in AI have been massive in the last few years, and we're starting to potentially see a turning point with DeepSeekv3 being created on a fraction of the cost and power of other models

DeepSeek (Chinese AI co) making it look easy today with an open weights release of a frontier-grade LLM trained on a joke of a budget (2048 GPUs for 2 months, $6M).

For reference, this level of capability is supposed to require clusters of closer to 16K GPUs...

https://techcrunch.com/2024/12/26/deepseeks-new-ai-model-appears-to-be-one-of-the-best-open-challengers-yet/

*This could turn out to be wrong hence why I'm keeping an eye on it **I'm absolutely certain a whole lot of execs are stunned right now they're spending billions when something that cost millions came up right next to them

[–] [email protected] 2 points 1 week ago* (last edited 1 week ago)

DeepSeek is really interesting! Leave it to China to solve some of AIs biggest problems - if we can have these models without needing football fields of compute powered by nuclear plants then it would totally change my perspective on the industry. As it is they're just too wasteful to justify their utility, but if DeepSeek just leapfrogged that problem then I have no complaints.