A hell of a lot of absolute slop as well… it’s really not at all obvious that it’s a net win in any way at this stage.
Actual real life case studies that people can point to where they deployed it and it all went great are actually very hard to come by. It was only the other week that there was that study that said the number was as low as I think they said like 5% or something.
No shortage of excitement though, but a lot of things seem to be riding on some big breakthroughs coming down the pipeline that may never arrive.
I feel like I am living in another reality when I hear about all these "studies" confirming LLMs have no value and how no one uses them. I am using LLMs all day. I have more ideas about how to use them than I have time to build out.
My day job and then family life consumes a lot of my time, but in 2025 so far I have shipped 8 personal tools and side-projects from zero, which is a record for me.(not to mention a lot of smaller stuff like custom browser extensions). That probably would have been 1 maybe 2 in a world without AI.
Actual real life case studies that people can point to where they deployed it and it all went great are actually very hard to come by. It was only the other week that there was that study that said the number was as low as I think they said like 5% or something.
No shortage of excitement though, but a lot of things seem to be riding on some big breakthroughs coming down the pipeline that may never arrive.