If that’s the only thing you care about you can DIY (or buy dedicated battery backup systems) with 10x more capacity for the same price. With heavier but cheaper Na-ion batteries in the horizon near future backup systems will be even better
Oh no, I have full creative control. I have to literally specify which fonts to use, where to place a shape etc. But I can ask it to pick a good color palette, put some kind of procedural texture etc. All of this would have taken me a day in photoshop, but with Claude I can give the instructions and let it use Python
If I tell an agent to trade stocks and maximize profits but the agent decides to hack something to achieve that am I on the hook for hacking? Genuine question
This site gave me the name “Greedy Bison”, but it didn’t say how many other “Greedy Bison” are out there. I’m guessing everyone using the very common device I’m using and using the very common browser that I’m using would get the same name as me
I’m not sure if this logic makes sense. They could easily just do the training in the backend with my prompt without showing me the autocomplete. The only difference is that I am slightly primed to the autocomplete, but the OP didn’t mention how that helps with feedback
US is an outlier, globally we are making great progress in using renewables and cutting down carbon emissions (still a long way to go)
Humans are making progress figuring out how to make space use sustainable as well! There are multiple companies out there working in de-orbiting satellites that have reached end of life. A single launch can possibly de-orbit 3-6 satellites. The capacity of a single one of these companies is enough to stabilize the number of space junk today
Top of Google’s page is just a worse version of AI, followed by three or four sponsored result, then SEO optimized crap and finally what I’m looking for is probably near bottom of the results page- I’m not surprised at young people ditching Google
Just because two graphs looks similar doesn’t mean they are “same”. Shouldn’t we subtract the random bias from the real world data and see if there is any slope remaining (or something similar)?
There’s argument to be made that SWE hasn’t been cracked either. The latest models are great at coding medium sized applications, but figuring out the requirements and consolidation of domain knowledge is something still lacking
While their conclusion is probably correct, I would have liked to seen the number of fatalities normalized by population, miles driven, number of pedestrian increase, speed limit change etc
Another outsider, but Krita seem to be doing really well keeping their UX up to date. To the point where I often use it for a lot of stuff meant to be done by GIMP
Who knows? The way things are going VCs could start asking claude to review someone’s online presence and Claude might decide karma is the best metric for that
Steam wishlist, itch.io number of views, YouTube views and now GitHub stars… I’m tired of all the gamification of creativity. Now if you’d upvote my comment I can get same karma please and thank you
Im not sure I follow. AI barely consumes energy compared to other industries and instead of focusing on the heavy hitters first wasting time on the climate impact on AI doesn’t seem useful