"Accelerationism" is an overdue corrective to years of gloom
thenewatlantis.com
thenewatlantis.com
Those "years of gloom" (which aren't very many years -- has everyone forgotten when the tech industry was widely seen in optimistic terms?) have been brought on by the behavior of the tech industry itself, in large part because of the misapplication of the idea "move fast and break things" (which is, unless I'm misunderstanding, the very essence of e/acc that this article discusses).
Our industry has been breaking a lot of things that people don't want broken, then tends to shame people for being upset about that. The problem isn't some inherent fear of tech itself, it's a (supportable) fear of the tech industry and what other things it may damage as time goes on.
If the industry wants to assuage these fears, the solution isn't to move even faster and break even more things, it's to start demonstrably acting in a way that doesn't threaten people and the things they hold dear.
People worry about AI paperclip maximizing, but Tech is already that in some ways (find or build moats, blitz scaling, no concerns for the harm incurred). It’s just fuzzy cohorts of tech workers and management doing the paperclip maximizing, for comp and shareholder value respectively. Not much different than AI reward functions.
People can only handle a limited amount of loss within a given period of time before they start pushing back hard against further loss and consider those causing them harm to be forces of evil.
There's also another factor that the tech industry is largely blind to: tech people tend to think that "we know best" and that pushing our ideas on the general public against their will is a Good Thing. But it's not a Good Thing, it's a Bad Thing.
Another thing we need to be doing is allying with the general public rather than dictating to them.
Disruption from tech advancement is caused by tech changes displacing existing industries and it hurts the people currently making money from those industries. But to be against that disruption you would have to believe that those people have some sort of right to make that money and continue doing the things that make them those profits when the public wants the more efficient tech. So really it's the anti tech people who are pushing things on the public.
E.g. people often complain about Amazon displacing small retailers, but really it's just that given the choice, most people choose Amazon.
That used to be true. Now, though, a very common thing I've noticed with people is that they use tech not because they want to or because it solves a problem for them, but because they are disadvantaged if they don't.
It's an important difference. If people willingly choose to use a thing, then they'll be inclined to think about it positively. If they use a thing because they feel they have no choice, then that thing is more likely to be viewed as adversarial, because it is.
I think that's largely where the tech industry has arrived at. Further, the tech industry shows little to no empathy to those whose lives are worse because of what it does.
Things start out good quality, high effort. Useful.
Then once they achieve a certain amount of inertia they start cutting stuff out. Adding new tiers of payment plans, or injecting advertising into existing plans. Lowering caps. Whatever else they can get away with to cut costs but keep your money.
But people have invested in them by that point. Invested enough that changing off is painful and potentially expensive. They want to barrier to leave to be high, even as they give you less reason to stay using their product.
What I see here is your basic Tragedy of the Commons at multiple levels. Consumers adopt the new thing to disadvantage their peers, who then have to do the same, to the detriment of all. And the vendors are doing the same thing.
This whole thread reminds me how much "tech" as a meme has come to conflate technology and business. People don't even seem to recognize that "move fast and break things" expressed a business philosophy, not some fundamental truth of technology, R&D, or science.
Edit: Sibling "enshittification" comment conveyed it better.
So yes, often people jump onto a hot new thing because all they can see are the benefits. The "buyer's remorse" doesn't come in until later, when the downsides become apparent. At that point, it's often too late and people are trapped. By design.
The tech industry counts on this effect, and doing that is one of the bad behaviors that encourages people to distrust the industry and become angry at it.
All I'm saying is that people are growing increasingly distrustful of, and angry at, our industry for really solid, rational reasons.
The most charitable interpretation I can think of for why we allow this to be is that the most visible part of our industry has become so insular and divorced from society in general that they can't even understand the anger or why it's rational. The least charitable interpretation I can think of is that they know perfectly well why people are getting mad and just don't care, because not caring increases short to medium term profitability.
We are. Constantly. We are, intellectually, in a tiny minority who find these things delightful and empowering. We assume that must also be good for everyone else. I was building electronics as a five year old when the other kids were playing outside and it thrilled me so much I assumed everyone else thought the same. They didn't and they don't. Maybe us nerds "took over the world", but as an adult I find almost everybody else (those we call normies) feel that digital technology is;
- something that happens to them
- is foisted upon them and they have no choice
- something they "have to trust"
> it gives people something they want
Have you considered that you really have no idea "what people want"? Neither do I, but I do know that and feel comfortable saying it. And I have done research and literally gone onto the streets interviewing lots of people to ask them. Most want what they think their friends want. Or the thing they already have with some new features. We tell them and they buy.
> to be against that disruption you would have to believe that those people have some sort of right to make that money and continue doing the things...
In a funny way they kind of do have that right. UDHR includes several aspects that can be taken as a "right to stability".
> it's the anti tech people
I don't encounter any "anti-tech people". Ever. I meet plenty folks who are anti-surveillance, or anti-authoritarian, or anti-asshole - against people forcing their technology on them - but I've never met anyone who thinks it's simply the fault of technology itself. You may be living in a bit of a bubble?
Software is pushing updates on us that nobody asked for in the exact moment we need them least, everything between the operating system to websites shove ads in our face, and features & "modern UI" that remove important information and options because designers thought a UI can't be confusing if it barely exists. Privacy violations pushed with countless license agreements and the daily "We value your privacy" popup that explains to you in what ways this sentence is a lie.
> it gives people something they want, not the other way around
I have never heard someone ask for slow and broken software, ads, tracking and other shady practices. They just have to live with it because what are they gonna do, not communicate with friends or file their taxes?
Also, plastic is looking far worse now than twenty years ago though I think the net is quite complicated and therefore ambiguous at the moment.
There is an ongoing discussion about “forever chemicals” and, again, not unambiguous but the balance seems to be tilting toward them being a bad idea.
I’m not personally seeing much of a dividend from nuclear weapons given how difficult nuclear power turns out to be under capitalism in practice. But I suppose it gets a pass because otherwise my father might have died in a land war.
Remains to be seen what the net will be on oil but I’ll happy speculate that if you consider a sufficiently long timeline that one turns out bad too.
I’m still pretty mad about the “food pyramid” but I can’t offer any particular study that tries to quantify its role in the decline of american health outcomes. Certainly modern food processing techniques look like a slow moving disaster but it’s really hard to sort our cause and effect.
Social media was neat for a few years but I would consider it a net negative.
I guess you’re right.
But that is a deliberately obtuse definition designed to justify any behavior.
If you let people continue their traditions, don't deliberately bankrupt them, and allow them make their own local laws, that's enough for most.
We were supposed to break; limits, barriers, status-quos, ossified ideas... Instead we broke; treasured social norms, privacy, mutual respect and dignity. There's a difference between benevolent innovation and reckless iconoclasm. I think it started the day Peter Thiel gave money to Mark Zuckerberg.
https://en.wikipedia.org/wiki/Meta_Platforms#History
>> '"On May 2, 2014, Zuckerberg announced that the company would be changing its internal motto from "Move fast and break things" to "Move fast with stable infrastructure".[40][41] The earlier motto had been described as Zuckerberg's "prime directive to his developers and team" in a 2009 interview in Business Insider, in which he also said, "Unless you are breaking stuff, you are not moving fast enough."[42]"'
ofc, with these stale game references around, I soon got reports of broken things: targeting was broken, pvp was broken, fade out animations were broken
A few hours later these issues were resolved. The players are used to these things happening sometimes. It's fine since the stakes are low. It's just a game after all. & being free, the active playerbase understands that they're QA
> The earlier motto had been described as Zuckerberg's "prime directive to his developers and team" in a 2009 interview in Business Insider, in which he also said, "Unless you are breaking stuff, you are not moving fast enough."
It’s about growth at all costs and then once Facebook got big enough they had to balance growth against other factors (+ the things people were doing that were causing breakages weren’t actually helping to grow).
Maybe true. But then if it's just about development it's a rather mundane old chestnut about reckless engineering versus good software engineering etc. Granted, that's a different discussion and we can see the tide turning now in terms of regulation and mandated software quality.
Sure, the Post-Office/Fujitsu scandal, Boeing etc, show how bad software actually ruins lives, but for the most-part the externality imposed by the reckless software engineer is measured in "hours of minor inconvenience".
That said.. I wonder if you did a ballpark calculation of how much harm lies behind the Google Graveyard [0], whether the cost of what is broken outweighs the benefits of it ever having been made?
One person’s recklessness is another person’s calculated risk. The consequences of FB engineering mistakes are minimal in both impact to customers and FB’s business. As FB scaled, the impact to individual people is still largely minimal (perhaps even beneficial) but the impact to their own business is larger and same for their customers if their ads aren’t getting eyeballs. So they shifted as big companies do. It’s kind of the best case of thoughtful risk taking - we’re rolling out a new system and we don’t know what could go wrong at scale and we put in monitoring of what we think we need. If there’s problems we’ll catch it with our monitoring/alerting and rollback or fix. You see the outages but not 99% of changes that go in without anything going wrong which lets the business resolve issues quickly and cheaply.
As for Boeing and Fujistsu, I’d say those are very different situations and aren’t an engineering problem nor do they indicate a move fast and break things mentality. As with many things like that, the engineering mistakes are a small detail within the overall larger picture of corruption. Boeing wanted to escape being classified as a new aircraft and met a perfect storm of skimping on hardware and corrupting the FAA through regulatory capture. I don’t fully understand Boeing’s role with the recent failures as a subcontractor is involved, but my hunch is that they’re nominally responsible for that subcontractor anyways. Same goes for Fujitsu - bad SW combined with an overly aggressive prosecution mandate and then cover ups around having made mistakes based on the assumption that the SW was correct rather than assuming new SW that hadn’t run anywhere before may contain bugs (not really sure whether Fujitsu hid the bugs or if politicians did or what happened but certainly the Post Office officials hid the reports of the auditors that found bugs in the sw and continued with prosecutions anyway).
Btw in engineering classes, all the large scale failures we were taught about involved some level of corruption or chain of mistakes. A contractor not conforming to the engineering specs to save on costs (valid optimization but should be done extra carefully), overlooking some kind of physical modeling that wasn’t considered industry standard yet, kickbacks, etc.
The difference then, morally, is whether the harms are sudden and catastrophic or accumulating, ongoing, repairable and so on. And what action is taken.
There's a lot about FB you say that I cannot agree with. I think Zuckerberg as a person was and remains naive. To be fair I don't think he ever could have foreseen/calculated the societal impact of social media. But as a company I think FB understood exactly what was happening and had hired minds politically and sociologically smart enough to see the unfolding "catastrophe" (Roger McNamee's words) - but they chose to cover it up and steer the course anyway.
That's the kind of recklessness I am talking about. That's not like Y2K or Mariner-I or any of those very costly outcome could have been prevented by a more thoughtful singular decision early in development.
You now have raised a totally separate issue about the overall strategy and business development of the company which you’d be right about - if it were required to have a PE license to run an engineering company, Zuckerberg would have to have had his PE license revoked and any PEs complicit in what they did with tuning for addictiveness should similarly be punished. But the lack of regulation in any engineering projects that don’t deal directly with human safety and how businesses are allowed to run is a political problem.
But on the issue I'm really concerned about, do you think "tuning for addictiveness" on a scale of about 3 billion users goes beyond mere recklessness, and what do we do about this "political problem" that such enormous diffuse harms are somehow not considered matters of "human safety" in engineering circles?
Is it time we formalised some broader harms?
I've been deeply critical of the appalling behaviour of the Post Office and Fujitsu in the Horizon scandal but there's a world of difference between this and the impact of Facebook in 2009. One had a foreseeable and foreseen impact on people's lives. The other was a social network competing with MySpace and looking for a way to monetise its popularity.
You're absolutely right there,
Frances Haugen's leaked internal communications showed incontrovertibly that internal Facebook research had long known teen girls had increased suicidal thoughts and obtained eating disorders. Facebook and Instagram products exploited teens with manipulative algorithms designed to amplify their insecurities, and that was documented. Yet they consistently chose to maximise growth rather than implement safeguards, and to actively bury the truth that their product caused deaths [0]. Similarly the Post Office had mountains of evidence that its software was ruining lives yet engaged in a protracted, active cover-up [1].
So, very similar.
But what's the "world of difference"?
> looking for a way to monetise its popularity.
That's a defence? You know what, that makes it worse. The Post Office were acting out of fear, whereas Facebook acted out of vanity and greed. The Post Office wanted to hide what had happened, whereas Facebook wanted to cloak ongoing misdeeds in order to continue. Simply despicable.
Way I see it - Facebook come out looking much, much worse.
[0] https://www.npr.org/2021/10/05/1043377310/facebook-whistlebl...
Their barbarous "barriers", "status quo", "ossified ideas"
vs.
Our blessed "privacy", "treasured social norms", "dignity"
Facebook was famously unit-test sparse, for example.
1: https://twitter.com/KelseyTuoc/status/1588231892792328192
They all do it. Doom gets more clicks than happy talk.
Sure it’s been “if it bleeds it leads” for a long time, but not until digital advertising has it been “if you’re not bleeding profusely 24/7 then you are going bankrupt.”
Taxis (incl Uber) are bigger now than they have ever been. Add in Uber Eats and co., and there’s so much new demand for similar services. Even Amazon flex has shown that car based services have unmet demand and utility.
The taxi industry should have been broken down. It was good for humans to have better access to the market, even if it wasn’t cancer treatment. There’s tons of money going to research cancer. New energy opportunities are a huge source of investment already, let’s not pretend the existence of Uber foreclosed the opportunity to cure cancer.
Gene drives, nuclear weapons, space organisms getting to earth, self-reflection on potential risks is not a sign of failure unless your only goal is to move. We only get to make some of these choices, as an entire global civilization, once.
AI is not exactly in that category but it's sure not a sign of failure that the people making it are actually considering the results before just making decisions with world-scale impacts. It slows things down but that is important when decisions only get to be made once and cannot be reversed. That was less the case with earlier tech, so I don't understand the surprise at the difference in scrutiny over time.
(my emph)
The root of this is a deflated concept of "progress". Progress is a vector. It has a magnitude and a direction. And it has a context that entirely defines its value. You would not be happy if the doctor told you you have a progressive illness. But "progress" gets used carelessly as a bare noun.
We can’t deal on an adult basis with children, no matter their age, how long do we have to let them stop us?
Tech has shown its cards and people hate it for a reason.
There is that. There hasn't been a must-have consumer electronics thing since the smartphone. 2019 was supposed to be the year of VR. Fail. 2023 was supposed to be the year of the metaverse. Fail. Internet of Things turned out to be a dud. Self-driving cars are still struggling. All those things actually work, just not well enough for wide deployment.
LLM-based AI has achieved automated blithering. It may be wrong, but it sounds convincing. We are now forced to realize that much human activity is no more than automated blithering. This is a big shakeup for society, especially the chattering classes.
It isn't possible to do what I wanted with the proposed command, but the hallucination helped me to Google a method that worked.
In my experience the only more or less reliable way to avoid hallucinations is to provide the right amount of quality information in the prompt and make sure the LLM uses that.
Now Google's Bard/Gemini on the other hand quite frequently makes up stuff. So for now I'm sticking to GPT 4 via Chatgpt plus subscription to augment my daily dev work.
So, to repeat the question:
Do you just run the command and hope? Or do you double-check using the manpage that it isn't going to do something drastic and unexpected?
- Make edits to a Latex file which would have taken me at least an hour longer to do.
- Reverse-compile a PDF into Latex from a mere copy-paste.
- Translate a travel diary from French into English.
- Ask conceptual questions in difficult areas of mathematics. It's unreliable but it often has a "germ" of an idea. This is backed up by the Fields medalist Terence Tao.
- Helped me tutor someone by giving model solutions to homework and exam problems when I wasn't sure.
- Write a browser extension to do certain content blocking/censoring that none of the programs on my computer could do. I've never written a browser extension before, and this would have taken me a day longer.
- Give me feedback on emails I wrote.
- Helped me deal with a medical emergency.
I've also used an LLM for some of the elements on that list, though it seems like I should use them more.
These days AIs/LLMs are still rarified air. People do use AIs built into a SaaS product (auto-summaries, etc.), but it's still the minority. Some others are becoming facile prompt jockeys. A rarer few experts will run their own models on local laptops and servers.
But it is intrinsically complex technology that most users don't really "grok" in terms of how it actually works. It is fundamentally very, very, very different than the spreadsheet. And its adoption will have natural limits and boundaries.
I know people who do office jobs unrelated to tech who have slashed their workloads in half using LLMs.
What do you mean complex technology? You just type plain English into a prompt; it can't get less complex. Have you seen how complicated spreadsheets are?
For my entire career in tech, anything that isn't the most extreme of optimistic view points has been relegated to unnecessary and destructive pessimism.
Tech can't stand "bad thoughts". We had a very necessary correction in this industry but people couldn't handle it and just completely lost their minds on AI.
I say this as someone who works in AI very close to the metal (i.e. not just calling an API all day). While these tools are very impressive, when I talk to people outside of my area it's like they've all taken obscene quantities of some new stimulant and can't even connect with reality anymore.
Even the AI "doomers" are, by my view, extreme optimistic because they don't see, or maybe just don't want to see, how much all of the public discourse on this subject is largely smoke and mirrors to drive yet another tech bubble.
Personally I think this is just mass hysteria created by the increasing awareness of the fundamentally unsustainable nature of industrial society. Reading this article is like listening to some hallucinating maniac on the streets of SF scream at you about things that really aren't there.
And the doomers just add on "and the week after that we'll all be dead!"
I'm personally not very impressed by the AI tools I've used. Sure they're a neat toy. They do seem to keep getting better. Maybe they'll be good enough one day for me to actually want to use them and feel it's a benefit.
Unless you explain this statement, most people here are likely to dismiss everything you have to say on the topic of AI.
Try asking chatgpt or Gemini about something complex that you know all about. You’ll likely notice some inaccuracies, or thinking one related subject is more important than something else. That’s not even scratching the surface of the weird things they do in the name of “safety” like refusing to do work, paying lip service to heterodox opinions, or interjecting hidden race/gender prompts to submodels.
It’s good at generalist information retrieval to a certain degree. But it’s basically like an overconfident college sophomore majoring in all subjects. Progressing past that point requires a completely different underlying approach to AI because you can’t just model text anymore to reason about new and unknown subjects. It’s not something we can tweak and iterate into in the near term.
This same story has recurred after every single ML advance from DL, to CNN + RNN/LSTM, to transformers.
> Try asking chatgpt or Gemini about something complex that you know all about. You’ll likely notice some inaccuracies, or thinking one related subject is more important than something else. That’s not even scratching the surface of the weird things they do in the name of “safety” like refusing to do work, paying lip service to heterodox opinions, or interjecting hidden race/gender prompts to submodels.
For sure.
On the other hand, I recently started to convert these AI hallucinations into a feature for me: it is like asking a person who is somewhat smart, but high on some hallucinogenic drug, on their opinion on a topic of your interest. Depending on the topic and your own intellectual openness, the result can be ... interesting and inspiring.
I asked it a question about Christianity and it stated things with the tone and certainty of a preacher.
Just gross. And worse to put something like that in the hands of billionaires.
I am somewhat of a doomer not because I think it will be the Terminator. More like AI theocracy here we go.
I've found two so far: the review summaries on Google Play are generally quite accurate, and much easier than scrolling through dozens of reviews, and the automatic meeting notes from Google Meet are great and mean that I don't have to take notes at a meeting anymore.
It did okay at finding and tabulating a list of local government websites, but had enough of an error rate (~10%) that I would've had to go through the whole list to verify its factualness, which defeats a lot of the time savings of using ChatGPT.
Beyond that: I tried ChatGPT vs. Google Search when I had what turned out to appendicitis, asking about symptoms, and eventually the 5th or so Google result convinced me to go in. If I had followed ChatGPT's "diagnosis", I would be dead. I've tried to have ChatGPT write code for me; it works for toy examples, but anything halfway complicated won't compile half the time, and it's very far from having maintainable structure or optimal performance. Basically works well if your idea of coding is copying StackOverflow posts, but that was never how I coded. I tried getting ChatGPT to write some newspaper articles for me; it created cogent text that didn't say anything. I did some better prompting, telling to incorporate some specific factual data - it did this well, but looking up the factual data is most of the task in the first place, and its accuracy wasn't high enough to automate this task with confidence.
Bard was utter crap at math. ChatGPT is better, but Wolfram Alpha or just a Google Search is better still.
In general, I've found LLMs to be very effective at spewing out crap. To be fair, most of the economy and public discourse involves spewing out crap these days, so to that extent it can automate a lot of people's jobs. But I've already found myself just withdrawing from public discourse as a result - I invest my time in my family and local community, and let the ad bots duke it out (while collecting a fat salary from one of the major beneficiaries of the ad fraud economy).
I agree they hallucinate and write bad code and whatever, but the fact that they work at all is just magical to me. GPT-4 is just an incredibly good, infinitely flexible, natural language interface. I feel like it's so good people don't even realize what it's doing. Like, it never makes a grammatical mistake! You can have totally natural conversations with it. It doesn't use hardcoded algorithms or English grammar references, it just speaks at a native level.
I don't think it needs to be concretely useful yet to be incredible. For anyone who's used Eliza, or talked to NPCs, or programmed a spellchecker or grammar checker, I think it should be obviously incredible already.
I'm not sold on it being a queryable knowledge store of all human information yet, but certainly it's laying the inevitable future of interacting with technology through natural language, as a translation layer.
An interface that's incredibly difficult to produce consistent output. As far as I know, we have not found a way to even make it do basic tasks parsed from natural language without a prohibitive-to-most-use-cases error rate. It's amazing that it can produce pretty believable looking text, but it's abundantly clear that there's no reasoning behind that text at all.
https://chat.openai.com/share/15b30c88-d21f-4ffe-8c15-5b444d...
I just ran a query for ChatGPT to recommend various databases for a specific use case. Out of the seven databases it recommended, only one was a database actually appropriate; one suggestion was marginally acceptable; three of the recommendations weren't even databases.
I then asked it to provide a list of the most important battles in North Africa prior to the entry of the United States into World War 2.
It gave me five answers. Three of which occurred after the entry of the United States into World War 2.
AIs provides extremely plausible answers. Sometimes it will actually generate correct, useful output; but you cannot yet rely on it for correctness.
https://chat.openai.com/share/43ebf64e-34ae-402b-a3ce-0787e2...
I wish someone would build a site that we share chatGPT4 outputs like the above.
I can't remember the last time I have actually learned something from a human on a message board like this compared to chatGPT4.
Talking to people like this is just 100% waste of time.
I'll agree that GPT4 has completely replaced Google, stackoverflow, etc for me.
The only time I use Google now is for janky more human like situations.
For example, today I had to transfer DC roles from a Windows 2012 R2 server to a new 2022. They only have one DC. And the old DC has a basically unused CA service set up.
ChatGPT would have me "fix" everything first, wheras, I did find a forum post with a situation almost identical to mine that helped me cowboy it rather than being overly meticulous.
There is still value to human experience. For now.
Also, what model did you use?
There is clearly significant value to this tech and I'm still dumbfounded how strongly some people try to deny it.
Is it possible GPT just thinks[0] that any answer stated confidently is preferable over not giving an answer?
Promise I'm not just being snarky, legitimate wonder!
[0]: I know it doesn't actually think, you know what I mean
You’ll find people who claim to have doubled their productivity from ChatGPT and people who think it’s useless here.
Bing Copilot can’t even answer simple questions about businesses in a particular city without lying. When confronted, it will apologize and then repeat the same text verbatim.
Zero reasoning happening.
In their defense, for most people that is what happened. Sure, they've been using "AI" tools of varying degree for a long time (spellcheck, language translation), but now they have totally free access to something that behaves sorta like the AIs they've seen on the popular Star Trek shows -- where, incidentally, the AIs were also imperfect.
[1] https://sts-news.medium.com/youre-doing-it-wrong-notes-on-cr...
You have to slay bad things to make a great company, and you have to identify them before they're chewing on something vital. Which means going out of your way to speculate about whether things are actually bad or just annoying.
It's hallucinogens. That's the drug of choice in tech. What that leads to, other than the obvious hallucinations, is an obsession with set and setting. Bad vibes will literally give you a bad trip and ruin your high. I've seen this with wealthy shroom heads over and over again. Their entire life becomes centered around making sure the set and setting are always perfect, which means any little bit of someone trying to talk some sense in to them, they get ignored.
Once you start to think through the kinds of behaviors that shroom addicts would start to engage in, especially if they had the wealth and resources to facilitate their addiction, you'll see it everywhere. It's not the typical "ruin your life" kind of addiction, but it's having an impact on what kind of ideas are allowed among the SV venture types.
To the degree Silicon Valley has a drug right now, it’s ketamine. (Before that, it was weed.)
I have a few friends in psychedelic assisted therapy and the effects I've noticed in them are the same effects I've noticed with regular psychedelic use outside of therapy: you don't identify the bad vibes and shy away from them, you end up making decisions that are uncomfortable in the sort term to improve things in the long term. Myself, I started going to college.
Maybe it works differently among wealthy people.
If you're in a cult of positivity, adding psychedelics to the mix is more likely to make you acutely away of the inauthenticity of the situation.
However, if you're in a positivity cult, and don't realize you're in a cult, and don't have someone guiding you to consider you might be in a cult, the shrooms are just as likely to make the cult seem like the most profound and important experience in your entire life.
If you are a lead in a company, you suddenly have a profound spiritual experience based around your ability to hire and fire people and tell them what to do and can use the drug to convince yourself that the ideas you are coming up with are the most profound thoughts a person has ever had. You won't even realize you're shutting out good ideas, because you have a messianic belief in AI or crypto or whatever the thing is, and you take the shrooms to reinforce that belief and you create an environment around yourself and put people around you that reinforce that belief.
It's a very different experience than going to therapy to work through your fear or depression.
Can you share some examples of what you mean by this? I've encountered people who are excited about "AI", for sure, but who are excited because a problem that has plagued them for years suddenly became solvable. Excited because the way they used to learn about things just suddenly changed. Excited because there's a better way to do research than to scour the crappy results provided by the big search engines.
There is absolutely an element of hype that far outstrips the reality, and after the single-function apps like "Summarize this paper for me" are monetized to death, we'll enter the trough of disillusionment, and some reality will set in. But there are absolutely transformative use cases unlocked by the latest generation of tools that are very real, exciting, and that enable a new generation of users to interact with computers in a way that was previously the stuff of science fiction.
I've spent much of my career working on boring enterprise tech, building software that's not flashy, but gets work done. The problem space is extremely large, and the solution space is extremely inadequate. It's this vast space of unsolved business problems that make some of this hype more real, IMO. LLMs in particular will be transformative for many of the problems the big enterprise platforms solve, and have the potential to solve some of the messiest parts of operating in that space.
I don't know that I agree with the "Years of gloom" characterization, and I agree that there are other underlying currents re: sustainability, but I can't help but feel this comment endorses a different problematic extreme. In the middle of the hysteria are real uses cases that will change how we interface with computers, and industries will be transformed/new categories created.
Anecdotally, GPT4 has completely changed how I approach research and troubleshooting and has been saving me many hours on a regular basis. Weird error message in a linux log file? I'll provide some context and paste the error, and a few minutes later the issue is solved based on a decently good answer.
Picking up a new language? I'll ask for example code to solve a very specific use case, and then keep returning when I run into errors. The rate of learning enabled this way is pretty remarkable.
If the only thing that we get out of this craze are some extremely high quality LLMs with an accurate understanding of most human knowledge, that alone is an incredible jump forward.
Nothing that goes against entropy is fundamentally sustainable. That doesn’t mean we can’t keep it going for time that is essentially infinite on human scales (unsustainable life has been kicking around on Earth for 3.7 billion years now). Defeatism is even more dangerous than hallucinating optimism.
Due to thermodynamics, this is a one-way process. In our case, it’s all fueled by the sun, which will keep burning for another few billion years.
But even that counts being not not truly sustainable, since one day it must end. The whole universe is destined to die a cold heat death where nothing at all happens anymore.
But that is a long time away and until then, we can build a beautiful civilization; we can learn and grow; and we can do so with nuclear & solar power.
People saying that we’re all as good as dead are technically correct, but in a very unhelpful way.
She previously wrote https://nadia.xyz/climate-tribes last year and there were critiques of her essay such as https://twitter.com/cognazor/status/1607509056499101696
It smells funny, because it is. People in e/acc claim to be "building" something fundamental that will change the world, change the future, etc. They are the heroes of the story.
But if you are actually building something, you are silent. Why? Because you're BUSY! Just like cryptocurrency: if crypto is going to revolutionize payments, and make things seamless and interoperable, and, and and... then why do you have to spend hours a day blabbering on about it?
The answer, of course, is because talking is easier than building, and with talking, it's all about you. Talking about cryptocurrency and then getting rich for talking is a lot easier than actually building it to make it useful. Likewise with this e/acc stuff. Most of the LLM startups are basically worthless appendages on the original tech. A pump and dump for VC money that just wants to be part of the next thing. Just like sh&tcoins were a worthless appendage on the original tech, and were similarly pumped.
I'll give them this, LLM are a lot more useful than crypto.
What's kind of sad about it all is that I actually think a proactive movement to build is a good idea. Let's build! How about with light rail, walkable buildings, bike lanes -- yknow, the things that actually improve people's lives? instead of "quantum qubit agentic neurographic models" or whatever else is being shilled?
The party is over. The cheap money is gone (it was wasted on a lot of bad ideas and bad implementations of good ideas).
The DJ has left. The coke guy is gone. And now you have to pay the carpet guy to get the stains out.
And it's just not plausible humanity could be that thorough. So, we might as well assume AI is not going to be that dangerous and move ahead.
(I am not an AI doomer, nor do I feel that nuclear weapons are not dangerous/should be less controlled)
It is not remotely feasible to control access to computing devices.
It's a way to process ideas you don't want to be true, sure, but it's not a sensible or cost-effective way to deal with potential threats.
(And yeah, you can argue AI x-risk isn't a potential threat because it's not real or whatever. That's entirely orthogonal to the "if it's true we're fucked so don't bother" line of argument.)
Global warming on the other hand is heavily grounded in model building - we're building models all the time, taking measurements, hypothesizing and then testing our models with reference to future data recovery. We have simulations of the effect of increased energy availability for specific climate systems, we have estimations of model accuracy over different length scales.
Where is this sort of analysis for AI safety? (if it exists I'm genuinely interested, but I just don't see it).
However, you don't have to work on the nastiest parts of the probability space, as no one has a perfect world model, and improving other scenarios should reduce the overall extinction probability estimate.
>“Accelerationism is unfortunately now just a buzzword,” sighed political scientist Samo Burja, referring to a related concept popularized around 2017.
Accelerationism is a political reaction that basically works like this: when faced with any kind of dilemma, choose the option that is the most progressively destructive. The idea is that society & institutions are so tainted, that the only way to fix it is to continue knocking down all those boundaries and Chesterton's Fence's in order to effect some kind of institutional collapse that's being headed toward anyway.
Why slowly implement policies that spell our doom when we can implement them fast - accelerate!
That's the idea. And they should have mentioned that in the article, because the contrast with the Globo Tech notion of "Effective Accelerationism" is just good irony.
It's about the gravity of the mob. Risk taking in business (which is now being called "acceleration" AFAICS) is about the courage not to think about "what everybody wants". I observe it's really hard for the SV tech mindset to escape that gravity. Downvote me all you like for saying such uncomfortable things.
You don't need to explain to someone why they want food. Of course they want food, they're hungry. You do need to explain to them why they want a pizza covered in gold leaf. You don't need to explain why they want a car: they want to get around in the United States, and a car is more or less mandatory: you need to explain why they want a $100,000 SUV that gets worse fuel economy than a comparable van while holding less cargo and has such terrible visibility there's a non-insignificant chance they will run over and kill one of their own children with the thing. You don't need to convince them they want a smart phone, you need to convince them they want a new smart phone that's 11% faster than the old one even though their current one works fine.
Tons and tons of business, not even remotely isolated to the tech sector, has nothing at all to do with meeting consumer demand or fulfilling wants in the market; it has to do with building slightly different versions of products that already exist, and then spending millions if not billions of dollars so you can scream at the market's ear as loud as possible until they think that voice is coming from inside their own heads, and they'll buy it to make it shut up. And then repeat.
If you're going to innovate, you ultimately have to go out on a limb and predict what people will want. That means high risk/high reward products not the kind of consistent reliable profits that Wall Street wants.
a tech subculture: creates its own acronym, posts a lot of self-referential coded messages, add exclamation marks, ignore those who decry it as annoying. no whining no complaining no critique. helps to get a tech alpha to engage/retweet you. tolerate a little discourse but mostly exclaim. make sure everyone knows you think X rulez. pro$it.
Personally, I don't think the widespread negative turn public opinion of the tech industry has taken has all that much to do with the tech. It's way more about the industry part. The business models of tech are defined by our current economic moment, one of rising inequality and governments that are increasingly complicit in supporting profit over human well-being. I don't personally think technology can be inherently good or bad, but what we do with it matters, and what we're doing right now is, in aggregate, reinstating a feudal society wherein the whims of powerful rentiers increasingly screw with people's lives. As far as I can tell EA and e/acc are just a distraction from this reality
The core belief of e/acc is that we can solve our problems not by austerity, but by growth. Instead of limiting births and limiting technological progress, we can instead find solutions to support more people and push humanity beyond its boundaries. Historically, it's the mindset that worked the best for countries, but it does have a limit.
Recently it got very culty and I since distanced myself from it. I don't think now is a good time to judge it since many people have their own understanding of it. You can just slap (e/acc) on your Twitter username and somehow become a voice for the "movement".
I don't think people should take it so seriously, it's just a way of thinking and everyone has their version of it. Just like effective altruism, people will use the label for their own agenda.
Honestly how I feel about anyone mentioning "accelerationism." It's not an ideology or movement, it's just a way to act sophisticated about trolling online.
This kind of straw-manning seems like the main game of e/acc types. Bin all criticism as gloom and fear, rather than actually engaging with the substance of it. Nobody (or at least virtually nobody) who criticizes SV culture is "afraid of the future", they're pissed off at a couple decades of shitty management grown fat, greedy, and stupid off ZIRP dynamics. Rather than face the obvious failures of the industry and its leaders, they'd rather pretend it's the kids who are wrong and blithely continue stepping on rakes.
I see similar parallels to the tech sector. Is the death of SV inevitable? If so, is the world better off if it dies quickly or slowly?
Kind of a tangent to the core article, but it seems to me that we are in a large-scale transition of society in many ways, and the redemocratization of technology is a critical aspect of a bright future..
You can justify horrible atrocities if you strongly believe in extreme outcomes.
Effective altruism, accelerationism, doomers... All of these mindsets are toxic for the same reason that extreme religious stances are toxic. Believing that the world is going to end or that it's going to reach salvation or whatever extreme outcome you feel like, and then using that to justify short-term atrocities is not the way.
Stop putting infinity in your forecasts. It breaks everything. We have been down this path before many times and it never ends well.
Yeah.. that’s the goal.
After such an atrocious track record of tech leaders being completely wrong about what "the next big thing" is, why should anyone trust these people and why are they so certain they're "accelerating" in the right direction?
Edit: I miss the era when tech companies treated their users as customers, not as marks to be manipulated for the purpose of profit maximization.
The original idea comes from Marxists advocating for the adoption of free market capitalism as a means of bringing about the alienation of the working class, which is a necessary precondition for socialist revolution.
The idea of effective accelerationism (which was a joke making fun of people like Musk and SBF) is the rapid, uncontrolled promotion of AI as a means of destroying the entire tech industry and all that surrounds it.
e/acc is about building fast and making changes that benefit humanity, while pushing useless bureaucracy away. Elon Musk himself is a proponent of e/acc. It has nothing to do with AI specifically, and especially nothing to do with destruction.
The core belief of e/acc is that we can solve our problems not by austerity, but by growth. Instead of limiting births and limiting technological progress, we can instead find solutions to support more people and push humanity beyond its boundaries.
> Effective accelerationism (e/acc) in a nutshell:
> Stop fighting the thermodynamic will of the universe
> You cannot stop the acceleration
> You might as well embrace it
> A C C E L E R A T E
https://beff.substack.com/p/notes-on-eacc-principles-and-ten...
where "the thermodynamic will of the universe" means endless darwinian struggle.
"This is just cope. What you’re describing is, is that reality is hard. Yes. If the thing we want is complicated and hard to get, the answer is not to pick something simple and easy and give ourselves a participation award. The answer is, well, we have to get stronger. We have to get better." -- Connor Leahy
"Discussing the risks and opportunities in front of us intelligently, e/accs believe, is a sign of a flourishing civil society"
Except e/acc has made a massive contribution to lowering the standard of discourse. Beff talked intelligently on Lex and they are much more reasonable on Twitter Spaces, but on the Twitter timeline itself 90% of their posts are some combination of trash/propaganda/insults.
"Rather, their moral vision is one where more people — including and especially those who consider themselves hands-off today — actively engage with emerging technology and identify concrete plans for its development and stewardship, rather than reflexively backing away from what they don’t understand"
Again, e/acc seems to be all about "build, build, build!" which stands in stark contrast to taking a step back and thinking carefully about the impacts of what you're doing before you do it.
"Discussing the risks and opportunities in front of us intelligently, e/accs believe, is a sign of a flourishing civil society."
Again, this isn't accurate. E/acc is very much not about balance and also very much not about about discussing risks intelligently vs. almost always criticising the people making the claims instead of engaging in discussion on a technical level.
...
Criticism aside, this article paints a picture of e/acc which, while not representative of the movement as it exists, is something which it could choose to grow into.
1. Things are just going to keep getting worse or stay the same. Preserve what we can, don't be a bad person, etc
2. It's up to us to make things better, our dreams a reality, etc
2 is a much more attractive general ideology for young people, and anyone not well situated financially.
If you are a techno-optimist, you should ask your AI to come up with something less associated with race war and neo-nazi ideology
In either case, seems very different than what proponents of e/acc are about, but in a cynical sense, not totally unrelated. The Wikipedia does include, "It has been regarded as an ideological spectrum divided into mutually contradictory left-wing and right-wing variants, both of which support the indefinite intensification of capitalism and its structures as well as the conditions for a technological singularity, a hypothetical point in time where technological growth becomes uncontrollable and irreversible."
You will have to be either a spoiled and bored (tech) billionaire to support it, or a clueless techy.
You forgot to control for "older people own more stocks because they spent more time saving for retirement".
do they know in the non IT realm, accelartionism refers to far right nazi groups believing the only way to bring back far right ideals is to accelerate conflicts with mainstream politics?
... and it's generally working, to the detriment of stability.
Nadia's article here seems like a much more reasonable version of this philosophy that embraces political solutions instead of naive pure tech ones.
By contrast, "Beff Jezos" (Guillaume Verdon) says he wants to be a "cultural engineer" [1, Lex podcast], but when recently challenged by Lex Friedman and Connor Leahy [2], he comes across like a dorm room libertarian who has an immature idea of what he's saying, not someone with something new to say, which I think is at least partially him trying to maintain a kayfabe [3] for the e/acc community (like a nichy techno Trump) and partially him actually being a true believer.
My critique [4] of e/acc, Jezos, Andreessen's thread is essentially that these philosophy-esque ideas are actually (unlike what they believe) not very ambitious in that they in practice advocate for: "don't touch the system, it works for me", and never mention (or seemingly talk about) how this tech can actually have massive impacts for regular people, or how we can use the gov't and broad prosperity and access to tech (and tech-enabled hyper-democracies) to further accelerate the best version of our vision for the future.
These guys are like the Underpants Gnomes of futurism:
Step 1: Build superhuman technology without any safeguards.
Step 2: ???
Step 3: Profit for me.
I at least expect "Profit for everyone" at step 3, but taking a try at Step 2 is what really would impress me.
[1] Beff on Lex: https://www.youtube.com/watch?v=8fEEbKJoNbU
[2] Beff v. Connor Leahy https://www.youtube.com/watch?v=0zxi0xSBOaQ
[3] https://en.wikipedia.org/wiki/Kayfabe
[4] https://twitter.com/NickPinkston/status/1714088788237180947
I am by no means an expert or huge fan of Land, but he's definitely more along the lines of, "the machines are going to eat everything and there's basically nothing you can do about it."
So, in some sense, this is just another round of "economic elites co-opting someone else's culture."
Edit: I did some searching on Twitter and found this great quote, which really does sum it up:
> E/ACC is a fitting end to Accelerationism. After having been passed around by dissident intellectuals and online deviants it can finally settle into retirement as another kitschy pastel MS Powerpoint Californian grindset aesthetic stripped of all its substantial insights.
If you read older posts, you'll find the earlier e/acc members got started with Fanged Noumena.
The fact that blame is required for the result of an election is kind of mind blowing. It's a throw away sentence yes, but it's also a very nice summation of the attitudes the Wests mandarins.
Trumps first victory happened because Obama threw the working class under the housing repossession bus. Trumps second victory will be because Biden threw the working class under the inflation bus.
Blaming tech is a very easy way of those people not having to look at the mirror and facing the desolation they have created.
But what gives those people the right to gamble with the lives of the rest of us in that way? The entire line of thinking is, in my view, not only horribly egotistic and authoritarian, but antagonistically so.