Tags: risk

Does AI benefit the world? – Chelsea Troy

Our ethical struggle with generative models derives in part from the fact that we…sort of can’t have them ethically, right now, to be honest. We have known how to build models like this for a long time, but we did not have the necessary volume of parseable data available until recently—and even then, to get it, companies have to plunder the internet. Sitting around and waiting for consent from all the parties that wrote on the internet over the past thirty years probably didn’t even cross Sam Altman’s mind.

On the environmental front, fans of generative model technology insist that eventually we’ll possess sufficiently efficient compute power to train and run these models without the massive carbon footprint. That is not the case at the moment, and we don’t have a concrete timeline for it. Again, wait around for a thing we don’t have yet doesn’t appeal to investors or executives.

AI Safety for Fleshy Humans: a whirlwind tour

This is a terrificly entertaining level-headed in-depth explanation of AI safety. By the end of this year, all three parts will be published; right now the first part is ready for you to read and enjoy.

This 3-part series is your one-stop-shop to understand the core ideas of AI & AI Safety — explained in a friendly, accessible, and slightly opinionated way!

( Related phrases: AI Risk, AI X-Risk, AI Alignment, AI Ethics, AI Not-Kill-Everyone-ism. There is no consensus on what these phrases do & don’t mean, so I’m just using “AI Safety” as a catch-all.)

Spring is arriving. It’s just taking its time.

There are little signs. Buds on the trees. The first asparagus of the year. Daffodils. Changing the clocks. A stretch in the evenings. But the weather remains, for the most part, chilly and grim.

Reality is refusing to behave like a fast-forward montage leading up to to a single day when you throw open the curtains and springtime is suddenly there in all its glory.

That’s okay. I can wait. I’ve had a lot of practice over the past three years. We all have. Staying home, biding time, saving lives.

But hunkering down during The Situation isn’t like taking shelter during an air raid. There isn’t a signal that sounds to indicate “all clear!” It’s more like going from Winter to Spring. It’s slow, almost impercetible. But it is happening.

I’ve noticed a subtle change in my risk assessment over the past few months. I still think about COVID-19. I still factor it into my calculations. But it’s no longer the first thing I think of.

That’s a subtle change. It doesn’t seem like that long ago when COVID was at the forefront of my mind, especially if I was weighing up an excursion. Is it worth going to that restaurant? How badly do I want to go to that gig? Should I go to that conference?

Now I find myself thinking of COVID as less of a factor in my decision-making. It’s still there, but it has slowly slipped down the ranking.

I know that other people feel differently. For some people, COVID slipped out of their minds long ago. For others, it’s still very much front and centre. There isn’t a consensus on how to evaluate the risks. Like I said:

It’s like when you’re driving and you think that everyone going faster than you is a maniac, and everyone going slower than you is an idiot.

COVID-19 isn’t going away. But perhaps The Situation is.

The Situation has been gradually fading away. There isn’t a single moment where, from one day to the next, we can say “this marks the point where The Situation ended.” Even if there were, it would be a different moment for everyone.

As of today, the COVID-19 app officially stops working. Perhaps today is as good a day as any to say Spring has arrived. The season of rebirth.

The Intelligence Illusion

Baldur has new book coming out:

The Intelligence Illusion is an exhaustively researched guide to the business risks of Generative AI.

I’ve had the opportunity to gather with my peers a few times over the past couple of months.

There was dConstruct, which I hosted. That was just lovely.

Then a few weeks ago, in spite of train strikes and travel snags, I went to Bristol to give a talk at Web Dev Conf, a really nice gathering.

This past weekend I was in London for State Of The Browser, this time as neither host nor speak, but as an attendee. It was really good!

I noticed something rather lovely. There was enough cross-over in the audiences for these events that I got to see some people more than once. That’s something that used to happen all the time but became very rare over the past two years because of The Situation.

None of the organisers of these events were pretending that Covid has gone away. Each event had different processes in place to mitigate risk. I wrote about the steps I took for dConstruct. For some people, those measures might seem to go too far. For other people, they don’t go far enough. This is a challenge that every in-person event is facing and from what I’ve seen, they’re all doing their level best.

None of these events were particularly large. Attendence was maybe somewhere between 100 and 200 people at each one. I know that there’s still a risk in any kind of indoor gathering but these events feel safer than the really big tech gatherings (like the one in Berlin where I got the ’rona—that was literally tens of thousands of people).

Anyway, all three events were thoroughly enjoyable. Partly that’s because the talks were good, but also because the socialising was really, really nice—all the nicer for being in relatively safe environments.

It’s not exactly an earth-shattering observation to point out that the social side of conferences is just as valuable as the content. But now that so many of us are working remotely, I feel like that aspect of in-person events has become even more important.

Or maybe I’m just appreciating that aspect of in-person events after spending such a long time with screen-mediated interactions only.

I had an extra long weekend recently. It was four days of being a culture vulture. It was also four days of ever-increasing risk assessment.

It began on Thursday morning with the first Salter Cane band practice in eighteen months. That was pretty safe—three of us in a room, reminding ourselves of how the songs go. I honestly thought it could’ve been a disaster and that I wouldn’t remember anything, but thanks to a little bit of last-minute revision the evening before, it actually went really well. And boy, did it feel good to plug in and play those songs again.

Later that day, Jessica went up to London. We spent that evening in the Royal Opera House, watching a ballet, The Dante Project. We wore masks. Not everyone else did.

The next day, the indoor gatherings continued. We went to the IMAX to see Dune. The opportunity was too good to pass up. It was wonderful! But again, while we wore masks for the duration, not everyone else did.

Still, I reckon the ventilation was reasonably good in both the Royal Opera House and the BFI’s IMAX cinema. But that evening we checked into the Clayton Crown Hotel in Cricklewood, venue for the Return To London Town festival of Irish traditional music.

That’s where we spent two days going to concerts, sessions, and workshops, all of them indoors. The music was great, and we had a lovely time, but I couldn’t help but feel a sense of nervousness throughout.

When we got back to Brighton, we both took lateral flow tests—thank goodness that these are freely available! We were both negative. We had dodged a viral bullet.

That was the last trip out of town we’ll be making for a while. But even for Brighton-based activities, this is the routine now: weigh up the risks, decide whether an activity is worth it, and if so, testing afterwards.

For example, the week after our trip to London, the Wednesday evening Irish music session at The Jolly Brewer pub started up again here in Brighton. It was one of the things I missed most during The Situation.

I wrote about this at the very start of the first lockdown:

I’m quite certain that one positive outcome of The Situation will be a new-found appreciation for activities we don’t have to do. I’m looking forward to sitting in a pub with a friend or two, or going to see a band, or a play or a film, and just thinking “this is nice.”

I certainly did find myself thinking “this is nice” during the session, which was as wonderful as I had remembered. But I was also thinking about ventilation, and distancing, and airflow. Like I said:

Risks. Benefits. Running the numbers. Making decisions. Trying to do the right thing. Trying to stay safe but also trying to live life.

The Situation persists. But it has changed. There are no longer any official restrictions to speak of here in England. Instead it’s down to everyone to figure out individually what’s right.

Everyone is evaluating the risks, making calculations and coming to different conclusions. It’s only natural that everyone—myself incuded—thinks they’ve found the Goldilocks zone. “That person is being far too blasé! And that person is being far too cautious! But I’m doing exactly the right thing.”

With that in mind, I’m trying not to be judgemental about the decisions other people are making (apart from the decsion not to get vaccinated—I’m judgemental as hell about those selfish assholes). For example, I wear a mask when I’m on public transport. Other people don’t. I try not to judge them (although really, how hard is it to wear a piece of cloth for the duration of a bus ride? …sorry, that’s judgy).

This tendency to comparison extends to the country level too. Right now England has the highest case numbers for COVID-19 in Europe. I look at Ireland with its magnificentally low levels of vaccine hesistancy and I’m jealous: why can’t we be like that? But then I look to the United States and think, yowzah!, things could be worse.

Jessica and I have made a couple of trips to London. One involved indoor entertainment—the long-delayed premier of Akram Khan’s Creature at Sadlers Wells. We wore masks. Not everyone did. I tried not to judge. Others would judge me for just being inside the building.

The other trip to London was a dog-sitting visit, hanging out with Cider who is a very good boy.

After each excursion like that, we do a lateral flow test. So far, so negative. Having access to free testing makes a big difference to making post-hoc evaluations of risks. It boggles my mind that testing is pricy luxury in the States (there I go again, making comparisons).

We’ve also started playing tunes at a session in our local pub. We make sure to test ourselves before going. Spending an evening in a pub—even a nice chilled-out pub on a Monday evening—is still a risk. But it’s worth it. Each time we go I think “this is nice.”

This isn’t the first time we’ve had to make risk assessments. I remember it was like this last year between the first and second wave. Can we sit outside a restaurant? Can we go see the new Christopher Nolan film?

But it feels different this time because of our vaccinations. I know the Delta variant has altered the game, but the stakes certainly aren’t the same as pre-vaccine times. So while I’m still trying to avoid catching this damned disease, I’m also trying not to let it rule every decision the way it did last year.

It’s a balancing act. It’s the same balancing act that everyone is undertaking. We’re all walking our own individual tightropes. But at least being vaccinated, the tightrope doesn’t feel quite so high off the ground.

I’m speaking at a conference in Lisbon next week. That means going to an airport. That means getting on a plane. That means spending time inside a conference venue.

But it also means I’ll be travelling to a country with a wonderfully high vaccination rate, where I’ll be able to enjoy the sunny weather and dine outdoors in the evening.

Risks. Benefits. Running the numbers. Making decisions. Trying to do the right thing. Trying to stay safe but also trying to live life.

The Situation persists. But it has changed. I look forward to it changing more. I’m in line to get a booster shot before the year is out. That’ll be another factor in my calculations.

I look forward to a time when I won’t have to keep making these calculations. That time isn’t here yet.

The Dangerous Ideas of “Longtermism” and “Existential Risk” ❧ Current Affairs

I should emphasize that rejecting longtermism does not mean that one must reject long-term thinking. You ought to care equally about people no matter when they exist, whether today, next year, or in a couple billion years henceforth. If we shouldn’t discriminate against people based on their spatial distance from us, we shouldn’t discriminate against them based on their temporal distance, either. Many of the problems we face today, such as climate change, will have devastating consequences for future generations hundreds or thousands of years in the future. That should matter. We should be willing to make sacrifices for their wellbeing, just as we make sacrifices for those alive today by donating to charities that fight global poverty. But this does not mean that one must genuflect before the altar of “future value” or “our potential,” understood in techno-Utopian terms of colonizing space, becoming posthuman, subjugating the natural world, maximizing economic productivity, and creating massive computer simulations stuffed with 10⁴⁵ digital beings.

Social Cooling

If you feel you are being watched, you change your behavior. Big Data is supercharging this effect.

Some interesting ideas, but the tone is so alarming as to render the message meaningless.

As our weaknesses are mapped, we are becoming too transparent. This is breeding a society where self-censorship and risk-aversion are the new normal.

I stopped reading at the point where the danger was compared to climate change.

Technical Credit by Chris Taylor

Riffing on an offhand comment I made about progressive enhancement being a form of “technical credit”, Chris dives deep into what exactly that means. There’s some really great thinking here.

With such a wide array of both expected and unexpected properties of the current technological revolution, building our systems in such a way to both be resilient to potential failures and benefit from unanticipated events surely is a no-brainer.

(Xrisk 101): Existential Risk for Interstellar Advocates | Heath Rezabek - Academia.edu

Exemplars proposing various solutions for the resilience of digital data and computation over long timeframes include the Internet Archive; redundantly distributed storage platforms such GlusterFS, LOCKSS, and BitTorrent Sync; and the Lunar supercomputer proposal of Ouliang Chang.

Each of these differs in its approach and its focus; yet each shares with Vessel and with one another a key understanding: The prospects of Earth-originating life in the future, whether vast or diminishing, depend upon our actions and our foresight in this current cultural moment of opportunity, agency, awareness, ability, capability, and willpower.