My current beliefs about AI timelines have made my life significantly worse. I find thoughts about the ever-shrinking timeline to AGI invading nearly every aspect of my life. Every choice now seems to be affected by the trajectory of this horrible technology.
Based on the responses in this post and others, many if not most people on this forum have been affected in a similar way. AI is the black hole swallowing all other considerations about the future.
Frankly, I feel a good deal of anger about the current state of affairs. All these otherwise thoughtful and nice people are working to build potentially world-ending technology as... (read 1428 more words →)
I'm confused how your estimate of the likelihood that Omicron evades vaccine protection could have gone up while your estimate of the likelihood that the previously immune are no longer protected against severe disease until they get a booster shot went down.
If vaccinated people are more likely to get infected, how is it less likely that they need an Omicron booster to protect against severe disease?
That's certainly the argument its proponents make, and perhaps it's true in some circumstances. But after looking at some of the research grants made to the Wuhan Institute of Virology in the years leading up to the outbreak, it's clear to me that there really is dangerous research out there whose theoretical benefits are vastly outweighed by the costs.
For example, there was some really dangerous gain-of-function research happening in a BSL-2 facility, which are just not designed to contain dangerous pathogens.
But we inherently need international agreements to deal with these problems. I believe the US banning gain of function research in 2015 was one of the driving factors that pushed the research overseas into Wuhan.
That website covariants.org is amazing. I've been looking for something like that for so long.
I think you are likely underweighting the possibility that Omicron could significantly decrease vaccine efficacy.
Omicron has 50 mutations including 30 in the spike protein. I believe Delta had 9. https://twitter.com/EricTopol/status/1464410605402480641
https://twitter.com/MichaelWorobey/status/1464325657861574659
"The B.1.1.529 (omicron) Spike looks a whole lot like a 'polymutant' Spike experimentally generated to evade antibody responses from infection or vaccination, as noted by @theodora_nyc
And at long last, here's the tweet I read two days ago and have been looking for ever since. https://twitter.com/theodora_nyc/status/1464257411582148608
> "I wish I was as optimistic @jbloom_lab Our polymutant spike has 20 aa substitutions and is almost completely resistant to neutralization by almost all vaccinated and convalescent plasma we tested. This new one has more in overlapping regions."
I... (read more)
The link is working for me, though the markdown you used isn't formatting the way it is supposed to.
So if I understand correctly, Moderna gave a third dose of either the B.1.351 variant shot or the original 1273 shot to patients and found that those given the B.1.351 shot had higher levels of neutralizing antibodies against the variant?
This seems to suggest that we could do strain specific boosters against COVID. I suspect that would be particularly true for Omicron since it has so much more antigenic drift than the other variants.
Probably for the same reason they downplayed the drop in vaccine efficacy from Delta: they didn't want to increase vaccine hesitancy.
What do you think of this new South African variant that seems to be massively out-competing Delta?
https://twitter.com/jburnmurdoch/status/1463956686075580421
Personally, I don't think I see that you can turn an ML system that has say 50-to-250 percent of a human's intelligence into an existential threat just by pushing the "turbo" button on the hardware. Which means that I'm kind of hoping nobody goes the "nuclear war" route in real life.
Isn't that part somewhat tautological? A sufficiently large group of humans is basically a superintelligence. We've basically terraformed the earth with cows and rice and cities and such.
A computer with >100% human intelligence would be incredibly economically valuable (automate every job that can be done remotely?), so it seems very likely that people would make huge numbers of copies if the cost of running one was less than compensation for a human doing the same job.
And that's basically your superintelligence: even if a low-level AGI can't directly self-improve (which seems somewhat doubtful since humans are currently improving computers at a reasonably fast rate), it could still reach superintelligence by being scaled from 1 to N.
Looks like a new even closer wild relative of SARS-CoV-2 has been found: https://www.spectator.co.uk/article/the-covid-lab-leak-theory-just-got-even-stronger
Seems to me like this evidence increases the odds that the origin of the pandemic was a lab in Wuhan. Anyone disagree?
Also, I really hate the FDA. Every month brings a new reason to hate them even more. The bureaucratic death machine continues to slowly roll over a few thousand new people every month.
I wrote this a year ago and was not really thinking about this topic as carefully and was feeling quite emotional about the lack of effort. At the time people mostly thought slowing down AI was impossible or undesirable for some good reasons and a lot of reasons that in hindsight looked pretty dumb.
I think a better strategy would look more like “require new systems guarantee a reasonable level of interpretability and pass a set of safety benchmarks”
And eventually, if you can actually convince enough people of the danger, there should be a hard cap on the amount of compute that can be used in training runs that decreases over time to compensate for algorithmic improvements.