site banner
PaperclipPerfector  4mo ago (text post)   25963 thread views

Culture War Roundup for the week of January 8, 2024

This weekly roundup thread is intended for all culture war posts. 'Culture war' is vaguely defined, but it basically means controversial issues that fall along set tribal lines. Arguments over culture war issues generate a lot of heat and little light, and few deeply entrenched people ever change their minds. This thread is for voicing opinions and analyzing the state of the discussion while trying to optimize for light over heat.

Optimistically, we think that engaging with people you disagree with is worth your time, and so is being nice! Pessimistically, there are many dynamics that can lead discussions on Culture War topics to become unproductive. There's a human tendency to divide along tribal lines, praising your ingroup and vilifying your outgroup - and if you think you find it easy to criticize your ingroup, then it may be that your outgroup is not who you think it is. Extremists with opposing positions can feed off each other, highlighting each other's worst points to justify their own angry rhetoric, which becomes in turn a new example of bad behavior for the other side to highlight.

We would like to avoid these negative dynamics. Accordingly, we ask that you do not use this thread for waging the Culture War. Examples of waging the Culture War:

  • Shaming.

  • Attempting to 'build consensus' or enforce ideological conformity.

  • Making sweeping generalizations to vilify a group you dislike.

  • Recruiting for a cause.

  • Posting links that could be summarized as 'Boo outgroup!' Basically, if your content is 'Can you believe what Those People did this week?' then you should either refrain from posting, or do some very patient work to contextualize and/or steel-man the relevant viewpoint.

In general, you should argue to understand, not to win. This thread is not territory to be claimed by one group or another; indeed, the aim is to have many different viewpoints represented here. Thus, we also ask that you follow some guidelines:

  • Speak plainly. Avoid sarcasm and mockery. When disagreeing with someone, state your objections explicitly.

  • Be as precise and charitable as you can. Don't paraphrase unflatteringly.

  • Don't imply that someone said something they did not say, even if you think it follows from what they said.

  • Write like everyone is reading and you want them to be included in the discussion.

On an ad hoc basis, the mods will try to compile a list of the best posts/comments from the previous week, posted in Quality Contribution threads and archived at /r/TheThread. You may nominate a comment for this list by clicking on 'report' at the bottom of the post and typing 'Actually a quality contribution' as the report reason.

7
Top Bottom Old Controversial Comments
Jump in the discussion.

No email address required.

Saw this linked elsewhere, and do we now have a better notion of what was going on inside OpenAI with the attempted ousting of Sam Altman? With Altman re-instated, and a new, Altman-friendly board in place, is this the kind of revision of the mission statement that was worrying the previous board?

I continue to believe the real danger from AI is not the AI itself, but the humans that use it, and "We can make zillions from fat, juicy, military contracts/Whoops, how we were to know that would happen?" seems like one of the failure modes that should concern the AI doomer set.

I don't see it as providing any new information: this was almost so certain as to be predetermined. The MIC is a lucrative and stable system to be integrated into, and the dynamics of capitalism were going to inevitably drive OAI into its arms. Little different from Google pulling out of China for being a totalitarian regime to, five years later, begging Daddy Xi to please let them make money in China.

Safetyists can draw some minimal level of comfort from the fact that OAI priorities will marginally shift from improving capabilities towards AGI to building tools that the MIC desires. More profits, less fundamental/deep research. (And I think that's genuinely good for safetyists: a smarter drone swarm is not going to destroy humanity.)

a smarter drone swarm is not going to destroy humanity

Not if you think the danger lies in the super-intelligent AI getting control of that smarter drone swarm to achieve its end. I don't believe in the super-intelligent AI, I do believe in ordinary dumb humans getting shiner, more destructive toys, and somebody presses the wrong button or is insufficiently clear about what or who the target should be, and then "oops" but it's too late then. 'How were we supposed to know that setting up autonomous killbots might come back to bite us in the ass? Sorry, widow of Mr. President, we never intended it to be our motorcade that got zapped, it was supposed to be the other guys over the border of the country our guy was visiting at the time'.

The future of Open Ai is much like Wolfram Alpha--something overhyped to change the world and initially really cool and useful, but now crap due to to extreme metering of computational power and paywalls. Expect the same here. The need to turn a profit from this will limit the power.

You know, I totally forgot about Wolfram Alpha. I remember it was really cool back several years when I last played around with it, but after I got through the math classes I had it help me with, it just flew out of my mind and was forgotten.

Disagree, GPT may be thoroughly spayed in the political sense but (as of now) this is still not enough to stop a slightly dedicated shitposter attacker, and while it has a long way to go in terms of cognition, what we have is already enough for many use cases. Cooding with it in particular is amazingly convenient, developers I know still have to wrangle it and correct its output but it is very tangibly helpful, and for me as a not-dev, being able to write simple scripts for work in 5-10 minutes of prompting instead of 1-2 hours of googling (especially if I'm a noob at the relevant language) is an absolute blessing. I imagine assorted wordcels feel the same way.

Don't think this is at all true. A lot of usecases are porn, and quite a few involve violence or 'being offensive' or racism, but the vast majority of llm usecases work almost as well as they would without censorship/safety training.

My hope is perhaps the technology is overhyped. Given the apparent change in just a few months about six months ago, I was expecting even bigger changes by now. Here’s hoping there is some kind of difficult technical issue.

I think there's more a staircase of jagged growth than a straight line up at the leading edge. You have leaps from GPT-2 to GPT-3 to GPT-3.5 to GPT-4. The gap between them can be years. GPT-3 was a cool toy but two years later GPT-3.5 changed the world. Even though there's been no qualitative change from GPT-4 last year, they've multiplied the context limit and added visual/internet browsing capabilities.

Why jagged? I reckon it takes time to process their gains understanding-wise, adopt new software improvements, get more training materials and buy new hardware. There's other stuff they do like censoring or optimizing . They're thinking 'do I really want to spend 100 million now and get a better version of GPT-4 or wait a bit longer for when my resources go further, once we've figured out X, Y and Z'. Training the models costs a lot of money and ties down a lot of compute. You wouldn't want to constantly be in training and making small improvments, missing out on making big improvements.

Just because nothing happened for the last few months, it doesn't mean that progress has stopped.

Maybe (though it’s really been more than a few months). We shall see. Perhaps it is merely copium.

The only interesting thing about the accelerationist vs safetyist wars is that for some reason the safetyists actually thought they were still going to matter once big money and the military-industrial complex decided what should be. EA hangers-on writing papers for their pet think tanks are just lucky they're unserious enough people to be sidelined through board wrangling and don't need to be thrown off any bridges.

The ‘safetyist’ guys are in many cases the very same engineers who demanded Altman’s reinstatement so they could get their Microsoft (or other) payout when it comes.

In truth, most people will abandon most principles when presented with a good chance of getting rich. And you can always justify it by telling yourself someone else was going to do it anyway, and at least this way you get paid.

I've definitely believed all along that once the money fountain got within their sights, all the "ha ha ha of course nothing will change/it changes" was the next step.

We solemnly swear we won't ever do anything even the teensiest bit naughty (unless it makes us a LOT of money).

I suppose it was just a bit eye-opening exactly how powerless the people who thought they were in charge really were, but yeah. AI danger is people, and people danger is greed, and "we can make tons of money off government contracts, don't let's be too fussy about which governments even" is always going to beat "AI can be an existential risk and we must be vewwy vewwy quiet when hunting wabbits", no matter how idealistic and "but look at all our high-quality technical papers full of the most jargoniest jargon!" you can pull out.

This is why I always thought the safety-first position— the people with power/money care more about increasing their wealth than in preserving humanity. And AI, if it actually works as promised, is a big, flashing “I win” button right in front of them. The only thing that might cause someone to consider blowing up an AI bank is if it belongs to a rival. The government wants it because it’s important to maintaining geopolitical status. The rich want it because the massive efficiency gains will put money in their pockets.

They've long been called useful idiots whose ideas around safety only really serve to establish and maintain control and do nothing to prevent evil uses.

I think this move vindicated that analysis.

'Evil uses'.

At the moment, LLMs couldn't plot their way out of a paper bag and fail at basic logic.

If we're talking 'evil' using LLMs to combat 'extremism' and 'misinformation' is both widely not seen as 'evil' and the most immediate use. They'll also be used to snoop through people's emails and highlight things cops could use.

Was any of that a use case OpenAI mission statement prohibited?

Image recognition seems good enough now though that killer drones that don't need an uplink and guidance to their own targets so they're immune to jamming are going to be fielded fairly soon. (<5 years).

But is OpenAI best at that?

I deliberately remained axiologically agnostic because what you think the machine should be restricted to do or not to isn't relevant to the fact that putting it behind a locked door and giving the key to the State and Corporations is never ever going to work.

They would have had a better chance putting it all in the hands of a single man. Organizations are structurally unable to stay mission focused. And only naive academics could believe otherwise to this day.

OpenAI bragging about its safety statement and mission statement which were going to make it the most ethicallest ever research company, don't be worried but just trust us guys. And now this.

I hate being proved correct about being cynical, I would have loved to be pleasantly surprised by "Huh, they actually do mean all the bumpf about safety and they won't cave in to the money fountain", but this is a fallen world after all.

What is this place?

This website is a place for people who want to move past shady thinking and test their ideas in a court of people who don't all share the same biases. Our goal is to optimize for light, not heat; this is a group effort, and all commentators are asked to do their part.

The weekly Culture War threads host the most controversial topics and are the most visible aspect of The Motte. However, many other topics are appropriate here. We encourage people to post anything related to science, politics, or philosophy; if in doubt, post!

Check out The Vault for an archive of old quality posts. You are encouraged to crosspost these elsewhere.


Why are you called The Motte?

A motte is a stone keep on a raised earthwork common in early medieval fortifications. More pertinently, it's an element in a rhetorical move called a "Motte-and-Bailey", originally identified by philosopher Nicholas Shackel. It describes the tendency in discourse for people to move from a controversial but high value claim to a defensible but less exciting one upon any resistance to the former. He likens this to the medieval fortification, where a desirable land (the bailey) is abandoned when in danger for the more easily defended motte. In Shackel's words, "The Motte represents the defensible but undesired propositions to which one retreats when hard pressed."

On The Motte, always attempt to remain inside your defensible territory, even if you are not being pressed.


New post guidelines

If you're posting something that isn't related to the culture war, we encourage you to post a thread for it. A submission statement is highly appreciated, but isn't necessary for text posts or links to largely-text posts such as blogs or news articles; if we're unsure of the value of your post, we might remove it until you add a submission statement. A submission statement is required for non-text sources (videos, podcasts, images).

Culture war posts go in the culture war thread; all links must either include a submission statement or significant commentary. Bare links without those will be removed.

If in doubt, please post it!


Rules


Recommended Posts And Communities

Recommended Realtime Chats