* Take a low margin, fragmented industry littered with failed businesses
* Then take some cheap but kind of profitable behavior in that industry (take local delivery over phone, send your lowest paid kitchen staffer to run it)
* Then layer on some of the most expensive labor on the planet (devs, PMs, tech company workers) to build a replacement for that behavior
* Then treat your lowest paid workers poorly, attracting regulatory scrutiny
Interesting that gwern predicted this as well yesterday
> Translation for the rest of us: "we need to fully privatize the OA subsidiary and turn it into a B-corp which can raise a lot more capital over the next decade, in order to achieve the goals of the nonprofit, because the chief threat is not anything like existential risk from autonomous agents in the next few years or arms races, but inadequate commercialization due to fundraising constraints".
> It's about laying the groundwork for the privatization and establishing rhetorical grounds for how the privatization of OA is consistent with the OA nonprofit's legally-required mission and fiduciary duties. Altman is not writing to anyone here, he is, among others, writing to the OA nonprofit board and to the judge next year.
Gwern's more novel prediction track record is calling everyone leaving from OpenAI (Mira was not expected) and general bullishness on scaling years ago. His post from 2 years ago (https://old.reddit.com/r/mlscaling/comments/uznkhw/gpt3_2nd_...) is mostly correct, though incorrectly believed large companies would not deploy user-facing LLMs (granted I think much of this is reasonably obvious?). And Gato2 seems to have never happened.
His overall predictions? I can find his prediction book which he heavily used in 2010 (https://predictionbook.com/users/gwern/page/2?filter=judged&... Brier score of 0.16 is quite good, but this isn't superforecaster level (there's people with Brier scores below 0.1 on that site).
Overall, I see no reason to believe Gwern's numbers over say the consensus prediction at metaculus, even though yes, I do love reading his analysis.
Sensible move since most of the competition is operating under a more normal corporate model. I think the non profit thing at this point might be considered a failed experiment.
It didn't really contain progress or experimentation. Lots of people are at this point using open source models independently from OpenAI. And a lot of those models aren't that far behind qualitatively from what OpenAI is doing. And several of their competitors are starting to compete at the same level; mostly under normal corporate governance.
So, OpenAI adjusting to that isn't that strange. It's also going to be interesting to see where the people that are leaving OpenAI are going to end up. My prediction is that they will mostly end up in a variety of AI startups with traditional VC funding and usual corporate legal entities. And mostly not running or setting up their own foundations.
This is it. Loss of trust and disagreements on money/equity usually lead to breakups like this. No one at the top level wants to be left out of the cash grab. Never underestimate how greed can compromise one’s morals.
My partner had a concussion many years ago, it was an eye-opening experience.
After months of mostly-useless conversations with doctors, neurologists, etc, we visited a sports medicine specialist who worked with snowboarders, skiers, etc, people who get concussions frequently.
The way the specialist described common concussion symptoms was really interesting: Effectively, your brain finds balance by using a combination of sight, touch (feet), and your inner ear. A concussion can impact the inner ear part of that equation, so your brain is overly reliant on sight and touch to compensate. This can cause all kinds of common concussion symptoms: Dizziness, sensitivity to screen time, etc.
Anyways, after giving us the rundown my partner was prompted to do a few simple exercises to test concussion symptoms. One of them was to stand on one leg and track a moving pen with her eyes. She'd done OK on some of the previous exercises but this one took her out, she lasted maybe 5 seconds and was completely exhausted and dizzy for the rest of the day because of it.
We ended up with a physical-therapy-like balance exercise plan that she stuck to regularly for a few months, and it ended up getting her to complete recovery.
Scotty, from Strange Parts, aka the youtube guy who showed the world that Apple could have included a headphone jack this whole time, went through a similar set of circumstances.
Like you said, extremely eye opening, and very good information to have. There is help available, and it likely won't come from a normal neurologist until their training catches up to the research.
People often seem to have a casual attitude to concussion; you read sports star X is out with a concussion and think little of it. Another word for concussion is "traumatic brain injury" and that tends to get it across better. I had a bad concussion a while ago and I still get occasional headaches, I had a tinnitus for months (if you are unlucky, it stays with you forever) and my sense of smell is permanently altered. I was ultimately lucky on that one too, enough force and you can permanently sever the connection to your olfactory receptors.
The neck is also surprisingly fairly involved in the balance system, I unfortunately have a lifelong inner ear deficit (deafness and vestibular) on one side and there have been times where neck tightness caused significant issues with my eye tracking (nystagmus). It’s really all connected, also salt intake, caffeine etc. can impact your balance / dizziness due to the pressure in your inner ear canals where the otoliths are swimming.
I once had a bad dream and violently shifted in bed hard enough to headbutt the connected nightstand next to my pillow (weird design). Hard enough to bleed.
I was dizzy for days whenever I laid down. It was as if I was spinning in an amusement park ride slowly.
I called an ex GF neuro radiologist, who after realizing I wasn't going to go in for any scans (no health insurance at the time), told me of a series of "brain/balance reset exercises" you can do. I did the exercises, moving my head in several positions in a particular order, and all symptoms went away.
More or less benign paroxysmal positional vertigo. The bang on your head dislodged crystals that form in the fluid in your ear’s semicircular canals. These crystals form with age and normally collect in some nook in the inner ear. When they are floating around in the fluid, they can strike the tiny hairs in your inner ear that indicate to your brain how your body is moving and how it is positioned. When fluid flows by the Haus and makes them sway, your brain interprets that as motion.
But when a crystal hits a hair, your brain interprets it as “I am suddenly moving about one axis at high angular velocity.” You become insanely dizzy. The nature of this phenomenon is that when you turn your head one way, you feel like you are falling forever in that direction.
Solution: the Epley maneuver. Turn your head quickly in the direction of the falling while you are sitting in a bed. As your head turns, fall backwards quickly on to the bed. This motion more or less tucks the crystals back into their niche so you can regain your sense of balance.
The brain is crazy. The inner ear is crazy. See an ENT specialist before you follow my dumbass version of doctor’s orders, but if you suffer from Bppv maybe this will be interesting information.
I studied neuroscience in undergrad and never learned about this. Thank you for sharing, I feel like I’ve experienced this phenomenon over the years and it might explain the changes in my behavior and erratic moods thereafter.
It wasn't exactly what you've described, but Im sure with enough googling someone could find all the various brain balance reset protocols and just try them all.
Downside: the Epley maneuver makes you feel so, so dizzy while it is resetting the crystals. Trying every one of these weird motion sequences one after another would probably give you some peculiar syndrome all on its own.
Like I said above consult a vestibular therapist, neurologist, or someone other than this dummy if you want real medical advice.
But it is real fascinating that this kind of physical manipulation is so effective.
Do you or anybody else have any resources to share regarding such training? I suffered a concussion 5 years ago and two more shortly after, and I still suffer from regular dizziness due to sunlight / busy environments and after too much screen time. It has taken away much of the joy of my twenties. Thank you.
Unfortunately I don't know anything online, but I would suggest seeking out sports medicine/physiotherapy folks who work with your local concussion-heavy athletes (snow sports, mountain biking, etc). Good luck!
The person we saw was a master of science in physical therapy, and specialized in/had personal experience with concussion rehab. They were based near a ski mountain, and thus a stream of athletes getting concussed regularly.
If you aren't near any skiing, I'd seek out sports medicine/physical therapy folks who work with your local concussion-heavy athletes. Mountain bikers, cyclists, maybe football/rugby, etc.
There is "vestibular therapist" specialty which is physical therapist that specializes in vestibular issues. I had one who helped with vertigo from BPPV and some of the recovery exercises sound similar.
Ie. Institute of Sport, Exercise and Health (ISEH) in London has both a neurologist (Richard) and a physio (Theo) specialised in concussion recovery, and working together. Been very useful for me (guessing you could do some assessments remotely).
Glad to hear your partner is better. My wife dealt with a serious brain injury in her early 20s (she has since fully recovered) and it’s always a concern of mine if she were to be hit in the head again.
> He told me to go to the FTC home page and look up the main phone number. “Now hang up the phone, and I will call you from that number right now.” I did as he said. The FTC number flashed on my screen, and I picked up. “How do I know you’re not just spoofing this?” I asked.
> “It’s a government number,” he said, almost indignant. “It cannot be spoofed.”
Completely insane that we continue to allow caller ID and number spoofing, it's so effective for these fraud scenarios.
I consider myself technically inclined, yet up until today I didn't realize numbers COULD be spoofed
One day a few months ago I woke up to a missed call from a verified number. I had been in a car crash the night before, and I was worried I missed a call from the driver's insurance company.
I called them back, and I was told that I was talking to a civil engineering firm; the receptionist was polite, but she sounded even more confused than I was. I had googled the number while I was on the phone, and yup, it belonged to a civie firm.
At the time I just assumed some engineer fat fingered my number by mistake, but I guess I missed a call from "Amazon" or "your insurance company" or some other nonsense. Funnily enough an insurance scam might have gotten me in the state I was in.
And, in fact, there are good reasons to spoof numbers. For example, a company may want all/many numbers to look like they come from a company's central exchange as opposed to an individual person's desk.
Don’t forget the kicker, that IAP at the time was unable to support more than a few thousand SKUs! And (iirc) that pricing, naming, etc for everything would’ve needed to be done through their atrocious web app.
This is not be the market success you’re thinking it is. The fast fashion industry relies on extremely dubious labor practices, massive amounts of pollution, and a culture of constant consumption and mostly disposable goods. 70lb of clothing are landfilled per person per year in the US alone (2018 numbers, today’s would certainly be higher).
The article’s comparison to gambling, and now outright use of gambling game mechanics by fast fashion co’s, is the most troubling part of the industry. It relies on getting people addicted to consumption, on ‘whale’ customers who buy constantly, on a lifestyle of single-use, disposable clothing (with massive negative externalities). Society, unfortunately, pays the cost.
I've had the pleasure of visiting a few cities/countries without much (if any) physical advertising/billboards, it's a change that's both incredibly jarring and incredibly pleasant!
At first there's a bit of an uncanny valley effect, particularly when driving around a city. Eventually that settles down, and the reduction in low-level visual stimulus is really appreciated. Coming back home, you might see the insane levels of visual pollution we deal with in a new light.
Not qualified to comment on the broader economic ramifications of this, but as a normal person navigating the world, it's pretty great!
The most annoying thing about them, that's seemed to increase over time, is that for 'premium' names the renewals are the same cost as 'buying' the domain.
At least if you manage to snag a .com/.net/etc, you'll have low yearly renewal fees.
* As the numbers here show, the high end of Manhattan rentals is very, very high. ~10%[0] of listed rentals at the moment are >$10k/mo. >$100k rentals are not unheard of. Looks like the highest market rental right now is asking $175k/mo.
* ~50%[1] of Manhattan's rental stock is rent stabilized, meaning the rent adjustment rates are set by the city. These units rent for considerably less than market-rate rentals.
* For a variety of reasons, rent-stabilized units often stay in the informal housing market (i.e. not on real estate listing services), which AFAIK these reports do not measure.
* There are a variety of cultural factors here that are a bit unusual for the US, including:
- It's fairly normal here for people with 'good' jobs to have roommates into their 30s+
- The informal housing market here is massive (e.g. I've lived here for 10+ years and have very few friends who've ended up in market-rate rentals)
- Manhattan builds very little new housing (e.g. zero units approved last month)
- It's Manhattan, so there's a LOT of money floating around here.
So yes, Manhattan is getting more expensive, but these market-rate rental reports are not the entire story.
[0] From Streeteasy, a local real estate listing service
these are not 'normal' rentals and imo should be excluded from stats, since they don't apply to typical families living year-to-year in nyc
a majority of those are short-term entire floor/home rentals, often rented out for less than a month at a time, to be used for production companies putting up A-listers for the length of a project, or even used as the set itself: sometimes the listing will say 'used for the filming of y reality show in 20xx!' like a selling point
Topical, someone just landed a Carbon Cub plane on a ~90ft helipad for a stunt with Red Bull a few days ago. The video[0] is insane, I've flown in STOL aircraft before but had no idea some were capable of landings like this.
Is the takeoff here less impressive than the Super Cub in the OP? It looks like the Red Bull plane falls off the side, but can trade altitude for airspeed until its back in normal non-stalled flight. The Super Cub starts on flat land and only ever goes up.
(And likewise, the landing is less dangerous than it looks because even if the pilot overshoots, he has plenty of time to get the plane back to powered flight after it falls off).
* Then take some cheap but kind of profitable behavior in that industry (take local delivery over phone, send your lowest paid kitchen staffer to run it)
* Then layer on some of the most expensive labor on the planet (devs, PMs, tech company workers) to build a replacement for that behavior
* Then treat your lowest paid workers poorly, attracting regulatory scrutiny