Even though I never really work on low level systems, by far the most useful courses I took at uni were on computer architecture and operating systems.
If you understand how computers work at a low level you can very quick grok almost any software dev concept.
Something I’ve noticed since starting to use AI to write all of my code: Its more difficult than ever to know when an app is feature complete. Its so easy to add more and more features
Are react server components the main way people build front ends in the React world?
I’m still using React + TanStack Router for single page applications, don’t know what the “industry standard” is these days
I think about that too. Obviously there is a very realistic future where the model providers increase token/subscription prices significantly.
But if there continue to be breakthroughs in hardware and AI research that can continue lowering the cost of inference and training, maybe the large labs never have to surge the prices.
I do wonder how the software profession is going to change over the coming years. At first people kept saying “the models aren’t good enough”, but now the models are pretty darn good.
I don’t think the whole profession is going to evaporate, but it’ll be interesting to see how it changes.
@coryalthoff See a lot of people saying selling, but I think engineering is the most important by far.
Its the most expensive skill to pay for, and without a product you don’t have anything to sell.
Hmm I guess I would just define things a little differently. I'd say getting leads is distribution, while sales is converting leads to users/customers. I guess it depends on a bunch of factors, whether your product has a free tier, if you're doing a PLG sales motion or more high touch, etc