Director Detection at SecureBio in Boston. Speaking for myself unless I say otherwise.
jefftk
Karma: 26,134
Thanks for asking about this! I just did a little research, and while it looks like the frequency range at 222nm would likely be sufficient for this reaction, there are three reasons why it’s not a concern in practice:
-
The UV light is not very strong: levels that would produce appreciable toxic gas would already be unsafe for human eyes and skin
-
Cleaning products almost always contain stabilizers to absorb free radicals
-
Human spaces have some ventilation (or we’d get too much CO2!) so any tiny amounts that were produced wouldn’t build up
-
I know this isn’t what you’re getting at, but reading your story made me think about how it is often the other way around with parents as they grow old and their children grow up. Watching your children, and everyone’s children, carry forward what you no longer can, and feeling pride in their accomplishments. And the children, now adults, playing harder and better by knowing they’re playing for everyone who came before them.
Yep! See paragraph 6.
Aww, thanks!
(I think of the “saint” framing as primarily being about sacrifice, but that’s not what we’re going for and someone looking over the material conditions of our life would probably conclude that we’re doing pretty well. Being able to give as described in the post is downstream from earning a lot of money as a software engineer in big tech and buying a house before the prices went up even more.)
Wouldn’t universities be even more incentivized to focus on fun over education? Students could take out loans, party for four years, declare bankruptcy, and then move on with their life. The bankruptcy falls off their credit history by 29.
I think it’s the other way: getting a loan to attend a primarily-fun school would be very hard, because the lender can see they’re unlikely to be repaid.
I wonder if your forecast is 10% because you’re much more pessimistic than I am about AI risk in general or because you have more aggressive AI timelines (or both)?
Both, I think, though looking at your numbers below I think it’s mostly timelines.
I think change might come very quickly, in a wide variety of ways.
if there really is a 10% chance that you’ll be dead two years from now due to an AI-caused catastrophe, it’s worth flagging that this means that AI represents about 95% of your mortality risk in the next two years
Yup. Worth taking seriously! I think it’s somewhat more than 95% of my risk, since I’m not likely to die of a drug overdose (no drugs at all) and don’t drive very much.
To answer this, what are your forecasts that you’ll be dead as a result of a global catastrophe by mid-2028/2030/2032/2035/2040? … 0.01X/0.05X/0.2X/0.6X
This is very off the cuff, but maybe 0.4X/0.8X/0.9X/0.95X? So 10%, 20%, 22%, 24% (and then 25%). Much more front-loaded than yours. My thinking is that AI is moving very quickly, and a large proportion of the risk is in worlds where we don’t have time to build defenses and general societal resilience.
in those 10% of scenarios where you’re imagining where we’re dead in two years due to an AI-caused catastrophe, what kind of AI-caused catastrophe are you imagining?
I think biorisk is a large proportion of the early risks, but accelerating things that don’t spread infectiously (drones, toxins) or massive societal collapse (no food) is also significant.
(2) What do you think is the difference in cost-effectiveness between the most cost-effective donation opportunities you can find and your current work?
Until recently I thought that SecureBio was close enough to my best donation option that, after taking into account the efficiencies of salary sacrifice, I should not take my full salary. I recently changed this after (a) SB started doing a lot better with fundraising and (b) with SB being a 501(c)(3) there’s a lot of money that can fund it which can’t fund other opportunities.
Or perhaps there is a different direct work job you could do that would be much higher impact than your work at SecureBio.
I think this is pretty unlikely, but feel free to try and make the case ;)
(1) Are your retirement savings invested optimally?
It’s mostly a Vanguard 401k, and I do have a lot of choice of funds. But I have it in total-market index funds for a combination of:
-
I’m skeptical of my ability to price stocks better than the market. Everyone can see the same returns that are pushing you towards investing in AI stocks.
-
Much of the benefit of the retirement savings is in worlds where AI slows down massively, in which case moving it all into AI stocks would have been quite a bad choice.
-
The cooking times in your first link feel borderline insane to me.
Agree! To be fair, though, this wasn’t one I was counting as a heresy. It’s pretty widely acknowledged that the high end of cooking time on boxes here will make your pasta badly mushy.
This step allows the sauce to get into the pasta, which won’t be possible in your plate.
Yup, I don’t want that. It’s neat that it’s possible to do this, but to my palate it makes the dish too homogenous.
I will spend more time with my child as a teenager than with a toddler.
A person is a toddler for ~2y and a teen for 7, but that doesn’t necessarily mean you’ll spend more time with them as a teen. The typical toddler needs a lot of supervision, and spends many hours a day with their parents. The typical teen is much more interested in spending time with friends.
(Ignoring short timelines here, which would push even more in this direction)
why not just play a clarinet then?
This should have better fingering than a clarinet while also having less keywork
Thanks for writing this post! I agree this is important.
it might help you to start behaving as if you’re being watched and things about you are more obvious than they once were
I’ve been banging this drum for a while:
2023-02 Preparing for Less Privacy
2011-04 Giving up on Privacy
A big one is that I’d need to set up a “please allow vibration” exception for every single app, and remember to do this as I added new ones.
Thanks! As
lsanderssays, in my case it’s about pressure, not the surface effect on the skin.
I’m still planning to use Ghiblification (in theory; I haven’t in about a year), but I didn’t think it was necessary here. These are pictures from a festival where it’s common for people to share pictures online after.
That would have large side effects: I don’t want DND always enabled on my phone.
Agreed! Some other of my kids posts along these lines:
https://www.jefftk.com/p/bets-bonds-and-kindergarteners
https://www.jefftk.com/p/perverse-independence-incentives
https://www.jefftk.com/p/how-to-parent-more-predictably
And maybe also:
If you want to spend money in an EA way, I think it’s pretty unlikely that feeding future looters is competitive with other EA options. For me, storing food is worth it because I value my and my family’s welfare more than arbitrary people.
Thanks for the reminder! “Check a box” cross posting is broken and I hadn’t done it manually yet.