Yes technically it is RAG, but a lot of the community is associating RAG with vector search specifically.
361 karma · joined March 14, 2007
One of our recent senior hires specifically wanted an in-office role and a lot of our junior engineers are hungry to learn from the seniors.
What I suspect is there will be a divide in the coming years and as long as companies are open about who they are, people can make the choice of where they want to work.
1. Chinese children prefer STEM content and the algorithm is providing that to them.
2. It's enforced by the Chinese government or someone who believes this kind of content will benefit the Chinese future.
In the case of #1, this is a cultural issue and we have no one to blame but ourselves.
In the case of #2, which I believe is what most folks who say this are suggesting, I can't imagine why children would then proceed to download the app, use it for hours a day, only to learn science and math. Sure, some may enjoy it (as some in the US would as well), but a vast majority of that market is going to reject this and delete it from their phone. In fact, we have this in the US -- we have educational TV shows, you can visit a library on your free time, etc, but kids don't do it-- because they're kids.
Getting the technology done even better means that Chromebooks and Androids (or even the browser) could be heavyweight gaming machines -- and thereby increasing those audiences.
As those audiences grow (and they're already big), having better hardware streamlined into Stadia means they immediately have insane distribution power. If you consider mobile gaming dominance in Asia, and the ability to release triple-A games immediately to a billion devices worldwide, that creates huge leverage over iOS.
Then as you said, GCP. I think someone with an idea and roadmap was there, but FAANG incentive programs don't reward people focused on the long game.
Still costs money, but if you're sending a lot of messages, this would substantially reduce the costs.
I find it hard to believe that Mixpanel has any kind of insight into this kind of information.
> Interactions, which occur when users click on a web link or advert inside the Facebook app, declined in seven of the last 12 months, with an average monthly fall of 2.6pc.
I think a more likely explanation of the decline is the following if all they're doing is tracking link clicks (I assume these are aggregate figures for all their clients):
https://www.forbes.com/sites/kathleenchaykowski/2018/01/11/f...
Though I'm not personally a regular user, from what I've observed, the AI is able to very, very efficiently resolve a significant number of requests because a lot of the messaging is the same: "what's the wifi?" "bring me towels" "can i get my car?" etc.
The application of NLP for a niche industry has felt particularly appropriate for a small startup where in contrast, doing generic NLP really feels like a boil the ocean type strategy that makes more sense to those who have the resources (google/apple).
To those who are testing: Akia shows up during the checkout process once you've entered payment and contact information.
Looking for feedback, questions, thoughts, or some first users on the product. :)
When I was at Facebook, this type of communication was extremely abundant and took place on Workplace. When you are pushed information, there is some expectation to respond (as highlighted by the author). However, Workplace's opt-in type of communication via feed didn't have this problem. And interestingly isn't really prevalent outside forums and groups (which I don't think are used commonly at companies).
I think splitting out that kind of behavior was beneficial, it wasn't email (which is a push that people ignore), but rather a subscribe where there was no stigma to be late, and people could take their own time to catch up on posts.
But the premise of the TV show emphasized the potential of the algorithm he developed. He had some serious IP on hand. Granted it's all fictional, and $10 million is not some trivial amount of money, but if you came up with an algorithm that proved P=NP (to put it in perspective), I'd say selling that IP for $10MM would be foolish.
And if I had that sort of IP, and if I was a person clever enough to come up with that in the first place, I'd take the $200k easy.
A practical generic solution, for example, that takes an extra 10 days to do, sounds fantastic from an engineering perspective-- we're trained to think this way. But on the business side of things, you may have a timeline for a feature to nab a client, or not enough budget for those extra 10 days, etc.
My opinion on these matters now generally falls into the it depends camp. Take his mentorship, because it sounds like he understands that every engineering solution is accompanied by a business solution and they must be agreeable for both parties (that said, he does seem a little jaded, so you don't have to take absolutely everything he says to heart-- there are companies who will give less weight to the business side of things, but the reality is that the consideration will always exist).
While I understand there are variables for every business, I literally have nothing to use as a gauge other than my own experiences.
So if you don't mind my asking, how many emails did you guys receive from being posted on HN?