there’s been a few times in the past ~30 years where you could essentially make a bet on an internet accelerated trend (e-commerce, cloud/SaaS, mobile etc. granted, you’d still have to be right about *which* companies would benefit & succeed, i.e. google vs yahoo) it seems like
consensus is building that AI/ML will enable the next generation of companies that will have a serious tailwind behind them (probably already underway to some degree) it seems like(from the little i know) theres an aspect of scale required to train models that benefits incumbents
and that it may be difficult for startups to be competitive w the large models developed by tech giants. i’m curious how people are thinking abt this phenomenon (and if you have any resources pls share) beyond the takes that are like “language models are going to disrupt google”
which may totally happen, but i’m more interested in the understanding the mechanics of how the coming years will play out wrt the value chain of like ok we need X GPU capacity (& other resources) to train a model that we can then sell for Y $ per unit of consumption or if you’re
the company paying for how much you need to use the model, how can you pass that through an existing product offering? (say if you’re microsoft or salesforce or whoever, do you start selling a version of your software that’s equipped w a LLM that you can charge a premium for?)
maybe to put it more simply at the risk of being cliché - how should one think abt buying the picks & shovels for the ensuing machine learning gold rush? and when u zoom out, who else stands to benefit & capture value from (what seems like) this period of accelerating innovation
Loading suggestions...