Depends on the job. My sister uses an AI (Gemini) so much that she has a subscription. Her husband also uses it frequently, and keeps it open in a tab at all times. He is a programmer, and she does marketing. I don't use an LLM for work, but very easily could if I wanted. All three of us mostly communicate with coworkers/partners/clients through email, so it's definitely doable.
The origin is really unclear. It's not certain which id employee added it to Quake 3 and where they got it from. John claims it wasn't him. Gary Tarolli, one of the founders of 3dfx, claims to remember tweaking the value of the hex constant to the value known today. But he says he didn't come up with the algorithm either. Maybe somebody else at 3dfx did, maybe he got it from somewhere else.
And at 3dfx he probably learned it from some ex-SGI guy, and that guy learned it doing his PhD at Stanford from someone who worked for ed Ed Catmull and so on. The lore goes deep with stuff like this. There's rarely just one author but many who improve the formula over time.
there's an unrecognized genius out there that realized you can just apply a bit flip to get that out there. That John Carmack hasn't claimed it as his invention when he could have, speaks volumes about his character.
Except he couldn't have even if he wanted to. There's always the risk some former 3dfx or SGI graybeard comes out with some paper binder with the original implementation, and calls you out on your bullshit if you do. When you're such a famous public figure no way you ever risk lying in public. Even if he were to getaway with it there was nothing for him to gain: he already has all the fame and money from honest work, there's no point risking it all to claim some piece of code.
Tesla promised self driving taxis and autonomous robots and cheaper new undisclosed products due to release soon, and investors bought it like they always do.
The stock market is a measure of what investors think the company will be worth over the next 10 years or so, essentially. Because if someone came along and wanted to buy the company, they’d typically pay around that amount. So take that amount (what the company is valued at) and divide by number of shares, there’s your share price. Like anything, it’s worth what someone will pay for it.
This is continuously adjusted as investor sentiment changes about where the company (and the economy at large) is going and what that “10 year valuation” is. A single quarterly result certainly could change that number if there’s a big surprise in there, but for the most part companies are good at predicting their revenue and expenses and that is priced into the stock already.
For Tesla, there was a wide expectation that they weren’t going to have a great quarter, it was pretty much known. So the stock was already significantly down before the earnings call. What was not expected was that Musk dedicated the company to building a cheaper car in the coming years. Investors thought this was a good move, stock goes up.
Now look at Meta, it was widely expected they’d have good results this quarter, so stock price is already up recently. What changed in this report is their expenses are much higher than what they previously estimated. So the stock goes down.
The actual quarterly results are a factor in the price but generally those are priced in already, unless you get some big surprise. What’s more important is the investor sentiment about whether the company is going in the right direction, which can change on a whim and for no reason at all.
It isn't that hard to figure out that one has been extraordinarily bullish and priced to perfection and one has been extraordinarily bearish as if they are going out of business.
There you go, spouting facts. Serves you right for being downvoted!
But really, this is exactly it folks. META has gone up like crazy over the past year. Do you think it will continue doing so ad infinitum?
Yea, my Irvine numbers are off. Torrance is becoming the next Irvine, due to a large Asian American community, decent schools, and relatively high paying employers nearby.
School district matters to Asian Americans, and Asian Americans skew in the highest income brackets in the US now, but we also prefer living in other AA neighborhoods because of cultural affinity.
Currently it only supports openAI. I'm looking into patterns for supporting local LLMs! I think the best approach for that might be to allow you to update the api endpoint to your own and assume you're using something that mirrors the structure of their API.
The main escape hatch for everything right now are custom nodes though. But then you'll need to bring your own templating pattern.
My goal is to build a single binary, local only command<->LLM interface. Chidori looks to overlap with that significantly, but of course i'd need custom LLM resolution.
