1,311 karma · joined July 5, 2017
It worked well on the management side of things too, but with an actual "changeover", as you'd have a Monday-Thursday supervisor and another Thursday-Sunday supervisor, with one day a week of overlap however that Sunday-Monday changeover didn't have an overlap (although there was a lot of OT on this job so it did frequently happen).
Scenario A, Sam Altman continues as CEO and the for profit arm of OpenAI continues to call the shots, growth and market share continue to be priority number one.
Scenario B, Sam Altman is fired, OpenAI non-profit board (re?)-asserts more control. Safety, alignment and other things like that take priority over growth and market share.
x% of Employees are Sam believers, and when the remaining x% of ambiguous/non-Sam believers realize the first x% might leave, their PSUs would be worth significantly less, so they sign the letter as well. There is also the peer pressure / fear of retribution factor as well once it becomes likely there is even a chance of Sam being reinstated.
Many employees choose scenario A because it is likely their "profit sharing units" will be worth more than with scenario B. There's a non-zero chance that OpenAI (the for profit arm) eventually joins the ranks of the "FAANG" companies. Those PSUs might be worth millions today, but in the future could be worth "fuck you" levels of money.
About ten years ago I was thinking of returning to university and was required to pass a high school level math exam to prove I still could do math as an adult (I couldn't, really!). Khan Academy was a pretty incredible resource. I was in a time crunch so I also hired a tutor for two hours a week and the combination worked really well for me. I could see swapping out a real life tutor with an LLM being pretty amazing and much more affordable!
I didn't end up returning to school, but now I've gotten the AI bug and need to upgrade my math skills, and I'm thinking that using a LLM to help learn about foundational math that applies to LLMs would be pretty neat.
There's a bunch of places in Europe now that have a no camping policy, but allow (or tolerate) "bivouacking", which originally meant sleeping on the ground without a shelter but now kind of includes small tents. I think it's the ideal tradeoff, it still allows long distance backpackers to do their thing while discouraging the "party camping" crowds.
Even pasting in https://google.com/search?q=apple+sauce&utm=14 to the url bar doesn't work. The only way I was able to get this to work was set DuckDuckGo (or anything non-google) as the default search engine, then use the extension "Smart Keyword Search" to setup a shortcut. I have !g <query> for normal google search and !gw <query> setup for google "web" search.
From the bottom up, there's a layer of bread, then a layer of ham, then another layer of bread, then a layer of cheese, then the top piece of bread. The ham and cheese are your conductive layers, one carrying current the other the ground.
When you first make the sandwich, everything is good because the ham and cheese are not touching each other. However, when you cut it in half, the bread, ham and cheese all smoosh together at the cut point and the ham and cheese electrically short out. Now, if you are super careful, you can probably go in with a super sharp knife, cut it again very carefully, and make it so the ham and cheese don't actually ever touch.
Apple (or whoever) purposely put a cut in the sandwich knowing the ham and cheese would touch, making the sandwich inedible. This YouTuber comes along with a really sharp knife and cuts the sandwich again making sure the ham and cheese never touch.
It just so happens when Apple made the cut, they didn't cut anything important, other than "smooshing" the layers (ie: they didn't cut through any traces or capacitors, etc).
You can actually read the photo from your passport with an Android phone. Some of the info in the machine readable area is required as a salt, so you can't do a remote scan of a closed passport.
With Llama, they now have an army of people hacking on the llama architecture, so even if they don't explicitly use any of llama-adjacent projects, there are tons and tons of optimizations and other techniques being discovered. Just making up numbers, but if they spend x billions on interference per year and the open source community comes up with a way to make inference even just a few more percent efficient, the costs of their open source efforts might be a drop in comparison.
For example, Zuck was on the Dwarkesh podcast recently and mentioned that open sourcing OCP (server/rack design) has saved them billions because the industry standardized their designs, driving down the price for them.
https://reactnative.dev/blog/2021/08/26/many-platform-vision
Side note: the 14" monitor is a game changer if your priorities are light weight and portability. Roughly half the weight of an iPad and even bigger and only like 60 euros. I went with the 2K resolution model and it's great price/quality ratio wise (using BetterDisplay to bump up into HiDPI mode).