@remidupre@mas.to titelbild
@remidupre@mas.to avatar

remidupre

@remidupre@mas.to

Dieses Profil is von einem föderierten Server und möglicherweise unvollständig. Auf der Original-Instanz anzeigen

fasterthanlime , an Random Englisch
@fasterthanlime@hachyderm.io avatar

One thing that is cool about LLM‘s right now is how there’s actual competition, the likes of which I haven’t seen in other fields lately.

I’ve been enjoying GPT-4o, now Claude Sonnet 3.5 is out and I can just switch to it with one line of config. Until next time.

remidupre ,
@remidupre@mas.to avatar

@fasterthanlime my issue about LLM is kot really whether they are relevant or not but how they so obviously included data they don't own to train their models.

fasterthanlime , an Random Englisch
@fasterthanlime@hachyderm.io avatar

I often think about "computer bad" and similar defeatism — so much is fixable via training, customization + just not buying garbage (which requires money, money solves a lot of problems!).

I'll never understand folks complaining about something they have the means to fix.

remidupre ,
@remidupre@mas.to avatar

@fasterthanlime also, French people are apparently known to complain a lot. The whole "the Olympics will surely fail" vibe seems like something that wouldn't happen elsewhere.

fasterthanlime , an Random Englisch
@fasterthanlime@hachyderm.io avatar

There’s still noticeable lag between the map being dragged and the paths following in Uber for iOS. What the fuck are they doing over there.

remidupre ,
@remidupre@mas.to avatar

@fasterthanlime I'm using an Android phone with MicroG (open source alternative to Google services) and it also appears to f**k it up. Most of the time the zoom level of the path and base layer are not correlated.

  • Alle
  • Abonniert
  • Moderiert
  • Favoriten
  • •
  • random
  • haupteingang
  • Alle Magazine
    •