Hacker News Reader

Comment by endisneigh | Hacker News Reader

I see a lot of positive sentiment and hype, but ultimately unless they own the phone ecosystem they will lose in the end, imho. In a year Apple and Google will trivially create something equivalent. Those who control the full stack (hardware, software and ecosystem) will be the true winners.

View on HN

True, this will not only be replicated by Google, Apple, Amazon and Facebook, but also by open-source. OpenAI has a short window of exclusivity. Nobody can afford to wait now, after reading the Sparks of Artificial General Intelligence paper I am convinced it is proto-AGI. Just read the math section, coding and tool use. I've read thousands of papers and never have I seen one like this.

https://arxiv.org/pdf/2303.12712.pdf

The question is how fast can you replicate it?

People will use the best solution. Chrome came after firefox and ie and opera and become more populare because it was better.

I am curious how Apple will approach it. They have historically valued 100% certainty with Siri above all else, even if it means having an extremely limited feature set. If there is even a tiny chance it might do the wrong thing, they don't even enable the capability.

I don't see how they can ignore this though. But at the same time it goes against all of Apple's culture to allow the kind of uncertainty that comes out of LLMs.

It’s not “trivial” because of the cost per query. As far as Google, it doesn’t even have access to the most valuable phone users without paying Apple $18B+ a year.

Something can be both expensive and trivial. If the market is huge they will bear the cost. The tech is well understood even now.

The parameter size will likely be an order of magnitude less for gpt4 level results in a few years

If the fixed cost was huge, you would have point. But the variable costs are also huge.

I’m sure the market is also huge for dollars sold for 95 cents.

it's really not that huge. unlike most companies, people are willing to pay if it's Apple. they can roll it into the iCloud subscription

That doesn’t help. You’re still making something that has a high variable cost into something that the user only is paying a fixed cost for.

I never said anything about the user only paying a fixed cost

So how well will it go over if the average consumer is nickel and dimed for every $x number of queries?

Not sure about "the tech is well understood" given the LLM itself is a black box in regards to how it works internally, even Microsoft researches admit it (Sparks of Artificial General Intelligence: Early experiments with GPT-4).

Reply on news.ycombinator.com