My partner and I have founded a company that uses custom AI models trained on research to (partially) automate the process of peer review and replication. We can identify mistakes and some types of fraud in research to aid reviewers as well as extract methods and equations from papers and automatically verify findings. If you know anything about the state of research right now, those are some incredibly large benefits.
I used to have a power shell script that a coworker gave me that would uninstall a huge number of services and apps on windows, change a bunch of config settings etc.
I've always wished there were a way to roll out a stripped windows release as an open source project without getting sued.
Yeah, it's both shitty and sometimes useful. It reminds me of an article I read once about implicit hierarchies- sometimes when organizations try to do away with traditional management, what they end up with instead is an unofficial and opaque control structure based on cliques and influence. In those cases it can be better for newcomers if there is an explicit set of rules and guidelines.
Not OP, but companies don't really care about people to that degree. They act for profit, or perceived profit, or to avoid a loss- someone that they know to be useful who is already familiar with the business is more valuable than an unknown.
I don't play games as much as I used to, but I had a lot of fun recently with Carrion - you play an alien hive mind eating scientists and fighting security through an Area 51 style facility.
The game Chex Quest, which was a total conversion of Doom for kids, used as an advertising campaign and included in cereal boxes. Incredibly well done game.
I had Kagi for a bit and enjoyed it, but I'm not sure I use search enough to justify the price tag.
I didn't know about the personalized SEO thing- I wonder if you could have a "default SEO rank" that would basically average all the specific uprank/downranks from other users. So power users tweak their algo, and everyone else gets the benefit of using that human feedback to improve their results.
You can train without a GPU, it just takes longer. More RAM and a better CPU will help up to a point. I don't think text generation is a particularly difficult task- you could probably do it with something like a Markov chain rather than an LLM if you don't care whether it's particularly coherent.
My partner and I have founded a company that uses custom AI models trained on research to (partially) automate the process of peer review and replication. We can identify mistakes and some types of fraud in research to aid reviewers as well as extract methods and equations from papers and automatically verify findings. If you know anything about the state of research right now, those are some incredibly large benefits.