|
|
|
|
|
by lateforwork
62 days ago
|
|
> There are plenty of Americans who can do these jobs. This thinking is wrong. For IT jobs, the work is pre-defined and you go find people who can do the job. For product development this is sometimes the case, but for truly innovative products, such as AI models, this is not the case. You have to hire the best in the world and give them the resources they need, as opposed to defining the project upfront and hiring people "who can do the job". |
|
I wrote my first neural net in the late 90s. Based on nothing but an old geocities post some rando put up about training a model to only unlock a pet door for their cat.
I implemented the same and it worked.
Where you see true innovation I see run of the mill. OpenAI, Google, etc are propping up data center rental business they came to rely on to titillate biology with whatever spaghetti that sticks. That's it.
The interesting science isn't happening anywhere close to big tech.
The mathematics of LLMs exists in textbooks from 1950s. Your entire comment chain here is little more than reciting propaganda.