So many people determined to get on the bandwagon--without considering whether they gain by doing so. The only time I've seen an AI that was actually of some use is Amazon's customer service bot. Basically everything it said was at least somewhat wrong, and it took considerably longer than a human would have. Their time saved, not total time saved.
I feel like I live in a weird other world. I find AI super useful. Then I read in forums like this how bad it is. I guess I am on the AI bandwagon. Obviously there are places where it still needs improvement but if you asked me a year ago if AI would be able to do what it does not I would have said no. So far I have really underestimated its potential.
I feel the same way, down to the "weird other world" feeling. AI has infinite patience, which means that you can ask and refine intricate questions. You can't always trust their answers, but you can't do that with people either. My favorite use for it is rubber ducky 2.0.
It is surprisingly good at rubber ducking and also that phase of code review which is more of a sanity check than a full review. I also like using it to build an index of source code in a new (to me) project. It is a tool and, like all tools, has things it’s better and worse at.
Except that due to a bias toward agreeability what actually happens is that it will subtly affirm all of your preconceived notions.
>"Lest we forget, MIT’s The GenAI Divide: State of AI in Business 2025 study found that 95% of AI projects fail to deliver measurable P&L impact."
Well we can pretty much ignore this after that line. It's completely meaningless to still be discussing AI capabilities pre-Claude Code and Opus 4.5.