Thing is that AI can help. My SO worked in a firm that does skill extraction from CVs and job ads. They do really cool stuff to match job ads with CVs using EU skill tags! It’s a really good tool to do specific things, so I really hate all the latest articles about LLMs.
I do find myself ignoring this kind of article, too, usually. I really enjoy discovering a totally new domain where the technology is implemented in a totally new way, going well beyond language applications even.
I dream of a ‘language’ model that specializes in general machine to machine communication. It almost surely exists already, but in my line of work, machine interfacing is an endless nightmare. A ‘protocol droid’ would be such a help.
Hey, but I managed to write software to calibrate µCT-scanners! That is clearly way more inspiring than all this fictional stuff. Right! Right. Right?
You bet your bippy that’s inspiring! An un-calibrated scanner just doesn’t hit the same way.
Based on the way specialized code is used, your calibration software will still be in use when they open the first scanning facility on the Moon.
Hope you accounted for the Y10K problem!
Thing is that AI can help. My SO worked in a firm that does skill extraction from CVs and job ads. They do really cool stuff to match job ads with CVs using EU skill tags! It’s a really good tool to do specific things, so I really hate all the latest articles about LLMs.
I do find myself ignoring this kind of article, too, usually. I really enjoy discovering a totally new domain where the technology is implemented in a totally new way, going well beyond language applications even.
I dream of a ‘language’ model that specializes in general machine to machine communication. It almost surely exists already, but in my line of work, machine interfacing is an endless nightmare. A ‘protocol droid’ would be such a help.