Hacker News Reader

Comment by summerlight | Hacker News Reader

In a PR perspective, at least AlphaGo did a prominent job for Google. It failed to keep its prestigious status but at least it's a leadership/product side issue, not DeepMind's.

View on HN

That stuff gave Google the appearance of "nerds playing with their toys." Sure it's worthwhile research, but it's not great how many things Google announced that close to 0 people could use. They're trying to be more user-oriented now, and because of Cloud, their users are businesses.

No. AlphaGo et al gave the appearance of "world class researchers achieving state of the art on problem classes that humans have been devoting lifetimes of study for thousands of years". the current desperate LLM scramble, including stuffing shitty, hallucinating, half-baked paraphrase spam to the top of every Google search result page, gives the appearance of "panicked flailing and tacit admission that high level decisionmaking has been captured by terrified MBA types"

There's a good long middle ground. Search results do seem desperate now, but they got into this position in the first place by failing to execute when they had the advantage. None of Google's customers have ever cared how well an AI can play Go. If there's one place the bragging could've translated to profits, it'd be Tensorflow dominance, which didn't happen.

That research has many applications besides just board games. Game theory is everywhere.

Also, there were serious mathematicians saying a few years prior that computers would never surpass humans in Go, at least during our lifetimes.

Yes it does, but that didn't translate to products, so again their customers don't care.

Reply on news.ycombinator.com