Occasionally reading what OSS AI gurus say, they definitely overhype their stuff constantly. The ones who make big claims and try to hype people up are often venture entrepreneur guys rather than actual ML engineers.
The open source folks I mostly keep an eye on are the ones who do actually code and train their own models. Some are entrepreneurs, but they know a decent amount. Not top engineers, but they seem to be able to curate datasets and train custom models.
There’s some wannabe script kiddies too, but once you lurk enough, you become aware of who are actually decent engineers (you’ll find some at Vector Institute and Jeremy Howard is pro- open source, for example). I wouldn’t totally discount them having an impact, even though some of them will overhype.
Occasionally reading what OSS AI gurus say, they definitely overhype their stuff constantly. The ones who make big claims and try to hype people up are often venture entrepreneur guys rather than actual ML engineers.
The open source folks I mostly keep an eye on are the ones who do actually code and train their own models. Some are entrepreneurs, but they know a decent amount. Not top engineers, but they seem to be able to curate datasets and train custom models.
There’s some wannabe script kiddies too, but once you lurk enough, you become aware of who are actually decent engineers (you’ll find some at Vector Institute and Jeremy Howard is pro- open source, for example). I wouldn’t totally discount them having an impact, even though some of them will overhype.