I'm fascinated by the large drop in England's overall suicide rates when the country switched from coal gas (containing carbon monoxide) to natural gas (much less lethal). It's pretty clear evidence that many suicides are impulsive acts, and simply making it a bit more difficult is sufficient to dissuade some people from trying.
I use a split keyboard with a magic trackpad in the middle between the two halves-- for me, it's so much more comfortable and ergonomic to move my hand toward the middle for mouse stuff, instead of off to the side, and using a split keyboard lets me keep my shoulders in a more natural position with my hands farther apart.
Here's a reference to Musk in 2019 saying that he would release robotaxis in 2020. "Musk pitched the robo-taxis as a way for Tesla owners to make money when they aren’t using their vehicles." https://edition.cnn.com/2019/04/22/tech/tesla-robotaxis/inde...
Same. Any time I visit a pharmacy counter for any reason, I add "and the largest box of sudafed you can legally sell me", to make sure I always have a stockpile on hand.
It's almost a shotgun house, which is weirdly common in some places despite being THE WORST architectural style for homes. I lived in one once, and having to walk through the bedroom to get from one side to the other is just as bad as it sounds. Bad for couples, even worse for roommates.
How did you choose between training a model from scratch vs using retrieval augmented generation with an existing off-the-shelf model? From what I've observed, RAG + off-the-shelf model seems to be the more common approach for use cases like "create LLM that answers questions about my company's internal documentation", particularly because the iteration/improvement cycle is much shorter-- it's much easier to iterate on RAG/prompts vs. training a whole new model to improve it.
(If the answer is "I just wanted to try training a whole new llm", I won't fault you for that! :) )
I like the idea that changing the display color was a must-have feature for the initial release. Shows a very solid understanding of the target demographic's user needs. (My 3yo loves any toy that includes a color picker!)
I spend about an hour a week at the grocery store. I bring my 3-year-old with me, and she brings her own little shopping cart and helps pick out what we buy. I think of it as a fun bonding activity and an opportunity for her to get exposure to a wide variety of food, rather than a tedious chore to be optimized. Sometimes she picks out weird fruits and then refuses to eat them, but overall I think it has helped broaden her food horizons.
Circa 2010-2014ish, I had the same Amazon delivery driver for several years, and it was awesome! It was just this one guy who delivered all the Amazon packages to my neighborhood. Same guy in the same truck every time, and he got to know my family and we would chat and he would help me with gardening and give me advice on how to prune my trees.
The biggest difference I've noticed is that the newer episodes are targeted to a younger demographic, which makes it seem very dumbed-down compared to the old stuff, but it means younger kids can keep up and learn from it. My 3yo learns a ton from the recent episodes, whereas the older sesame streets are just a little too complex for her to follow. I'll try the older ones on her once she ages out of the recent ones.
I'm so ready for checks to die, but somehow I always seem to have someone tech-averse that I need to pay monthly. I briefly got to stop writing monthly checks when my dog walker started accepting Zelle, but now my preschool will only take checks, so I'm back to writing a check every month.
You're certainly correct that software companies should do a lot more year+ A/B tests. You can learn really interesting things from it that a shorter test won't capture. I know of this one: https://medium.com/@AnalyticsAtMeta/notifications-why-less-i...
This article is magnificent nerd bait-- there is no possible ranked list that anyone would be satisfied with. You can't compare Totoro and Grave of the Fireflies, it's like asking "which is heavier, blue or pi?"
His kids book "Every night is pizza night" is also great. I've read several "get your kid to try new foods" books to my child, but that one actually got her excited to try red beans and tagine and dumplings and green pozole soup, and now those are all foods she will happily ask for!
I don't anticipate everyone wanting to do this, and I definitely don't picture anyone wanting to do "work" for 40 hours a week in a world where work is optional. I would think some people will want to do a few hours a week of something in the category I'm describing, and many people will elect other forms of meaning in their life (maybe they love whittling driftwood or golfing or just want to watch AI-generated movies all day. That's fine.) The whole point is we'll have options, so there will be some supply/demand curve for meaningful human interaction that will find equilibrium.
If they don't want to, that's fine with me. Just offering some options for people (like me) who still want a meaningful way to contribute to society, in a world where robots do most of our current jobs.
Instead of telling people to play video games or go kayaking, I hope we can ask them to go play checkers with the elderly, read to children, hold hands with dying people in hospice, and mentor the formerly-incarcerated. These are "jobs" that I would argue robots cannot meaningfully fill, but which society would benefit from a lot more of.
It feels like we could cut out the middleman by giving people money for doing nothing, and letting the robots do the dehumanizing jobs. I'm not sure why people think it's morally superior to force people to do useless stuff to get the money.
In what field do tenured professors take 3-4 months of holiday per year? In the hard sciences (US, R1 research universities), I saw tenured professors doing long hours of research all year, and treating summer as a time when they could focus more-fully on research and grant-writing with less of those pesky classes getting in the way.
I found therapy really useful for having a professional to guide me through the self-discovery process to identify some things I really want in life-- and more important, some things that aren't a priority for me (which I previously thought I really wanted). Figuring out what you DON'T want is at least as important as figuring out what you want.
As a result, I changed my career path (slightly! not even a whole blow-up-your-life thing) a year ago and have been happier since.
I also really liked the book Four Thousand Weeks, which is all about figuring out how you want to spend your finite lifetime.
I agree with both sides here. :) DS should own experimentation, AND engineers should be able to run a majority of experiments independently.
As a data scientist at a "blue chip company", my team owns experimentation, but that doesn't mean we run all the experiments. Our role is to create guidelines, processes, and tooling so that engineers can run their own experiments independently most of the time. Part of that is also helping engineers recognize when they're dealing with a difficult/complex/unusual case where they should bring DS in for more bespoke hands-on support. We probably only look at <10% of experiments (either in the setup or results phase or both), because engineers/PMs are able to set up, run, and draw conclusions from most of the experiments without needing us.
I wonder if some of the bad Amazon experiences are due to counterfeit/fakes, and/or the product lines that are made in china? Their products vary in materials and country of manufacture. The ones I like are made of Vitrelle, come directly from Corelle, and it's one of the product lines they still make in New York: https://www.corelle.com/product/winter-frost-white-78-piece-...
I think they're all (even the vintage ones) subject to risk of shattering into pointy shards if they break, though. They're not magic. :)
I use borosilicate lab beakers in the kitchen. The main disadvantage is that they don't have handles so you'll need a pot holder to use them for hot things.
Everyone who wants one already has one, and it's not the kind of thing you need to upgrade every year. I'll keep using my Instant Pot for the next 10 years, but that won't help Instant Brands make money.