site banner
PaperclipPerfector  1yr ago (text post)   30032 thread views

Culture War Roundup for the week of May 8, 2023

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.

5
Top Bottom Old Controversial Comments
Jump in the discussion.

No email address required.

I know some of you may be sick of "wow, look at this cool thing AI can do, how eerie is that?" posts, but I can't help but be blown away by some of the applications of this technology.

I play in a band, and we upload our music to streaming platforms using a company called DistroKid. When you're uploading music, they strongly encourage you to submit the lyrics for the song at the same time for SEO reasons (and so that the lyrics will appear onscreen if someone shares the song on Instagram). It's a little textbox where you paste your lyrics in, then you're done.

I went to upload a song today, and found that they've added a new feature. After uploading the audio file, you can get an AI to try to transcribe the lyrics rather than typing them out manually.

It nailed it. The song is 215 words long, and the AI only made about 6 mistakes i.e. 97% accuracy. For reference, this is a noisy post-hardcore song with layers of shoegazey guitars, feedback and pounding drums. What's more, it achieved this 97% accuracy in a matter of seconds, maybe 5% of the total runtime of the song itself.

A few years ago I used a voice transcription service called Happy Scribe, which achieved comparable levels of accuracy - but only for plain speech recorded in a quiet environment with no background ambiance. If you record speech without a directional mic in an environment with a lot of background noise (e.g. a café), Happy Scribe was pretty much useless and I had to transcribe everything the old-fashioned way.

But now AI can transcribe words near-perfectly from lyrics with a musical accompaniment? That's insanity. I can't imagine how anyone will be able to find work as a transcriber for any major language a year from now.

Checked three random songs with whisper.cpp (ggml-medium.bin -l en): Grimes' 4ÆM, Kanye's Champion, and Juno Reactor's Angels and Men. Kanye got parsed ≈perfectly, the other two gave me lines of [MUSIC] [MUSIC] [MUSIC] and (dramatic music) (dramatic music) (dramatic music), which is fair enough but very helpful. On this note, it often interprets silence at the end of my voice notes into nonsense like «subtitle editor Lena Yeterenko, corrector T. Ilyin» or «[popular streamer] is done for today, like & subscribe», which I vaingloriously attribute to my professional diction and sexy voice despite it really being just an archetypal case of overfitting. Nevertheless, I think with a little tinkering this is all doable with basic Whisper and that's probably what runs on the backend of your service. We're 99% of the way to solving transcription.

And a bunch of other tasks, to put it mildly. Ethan Mollick recently wrote a nice piece on ChatGPT+Code Interpreter: It is starting to get strange. (My somewhat-related earlier comment).

Code Interpreter is GPT-4 with three new capabilities: the AI can read files you upload (up to 100MB), it can let you download files, and it lets the AI run its own Python code. This may not seem like a huge advance, but, in practice, it is pretty stunning.

Lets take an example: I am writing a blog post about how amazing ChatGPT is at working with code right now. I would like you to create the perfect illustration, a GIF using Python, that represents this ability. Decide what an appropriate amazing GIF would be, then figure out how to create it and let me download it. After its first attempt, I encouraged it to do something even more creative. It decided on a strategy, wrote software to enact its strategy given the constraints on its tools, executed the code, and gave me a download link to a GIF.

…So the AI shows genuine creativity in problem solving. That seems like a big deal, but not actually the big deal I want to discuss. I want to show you that Code Interpreter has turned GPT into a first-rate data analyst. Not a data analysis tool, but a data analyst. It is capable of independently looking at a dataset, figuring out what is interesting, developing an analytical strategy, cleaning data, testing its strategy, adjusting to errors, and offering advice based on its results.

An example: I uploaded a Excel file, without providing any context, and asked three questions: "Can you do visualizations & descriptive analyses to help me understand the data? "Can you try regressions and look for patterns?" "Can you run regression diagnostics?" It did it all, interpreting the data and doing all of the work - a small sample of which is below.

etc. I'm sure that a great deal of blue-collar work of this kind, work that is well paid for, is actually not very valuable, if not outright bullshit in the style of that Bernanke&Krugman joke: routine extraction of so-what observations and weaving them into platitudinous narratives for non-actionable powerpoint presentations for people who make presentations of their own to drive up the company valuation by justifying the foundation of the whole house of cards.

But some of it is valuable – immensely so. Markets have not yet recognized that we have an alien artifact on our hands. It may create unemployment, albeit that's not a given; it definitely is able to add value comparable with the bigger breakthroughs of the 20th century. And this is just about GPT-4-tier systems.

Yud of course prophesied that AIs won't be very useful until very late in the game where they rapidly become deadly (I can't be bothered right now to hunt for specific quotes). I think it's in clear sight that even just continuation of this line of tools will be sufficient to accelerate basic research to the point that we get either controlled nanotechnology or proof of its unfeasibility before evil agentic AI pulls itself up by its tentacles and killseveryone. Not to mention the possibility to maintain decently efficient local supply chains and other infrastructures even for small off-grid communities, obviating major value propositions of a nation state.

But the more likely scenario, IMO, is that «evidence» from recent and upcoming Hollywood flicks, Yud's podcasts/TED talks and obsessive fearmongering of mentally unstable people will be used to justify severe «compute governance». The Elders seem to be of the mind that useful artificial intelligence is no toy for children. You've got your guns, you've got your trucks, you've got your free press and your flags; and be happy with those inconsequential obsolete freedoms. LARP as a Butlerian Jihadist (while swiping vigorously on your personal data harvester device) or WH40K's loyal citizen (temporarily embarrassed by an onslaught of Slaanesh's forces) if that's your flavor of cope – just do not aim above your station. No «digital god» for you, little man; pray to your paper gods.

More than songs, I just want podcasts to be transcribed, as a rule. I'm far more into reading (and writing, why else would I even write here?) stuff than listening to it, even listening to it at 1.5x speed or whatever, especially when it comes to English. Sure, there are podcast transcription services, but they require me to download them and then plonk them to the program one at a time for transcription, and the results are still so-so.

Have you tried AssemblyAI? They have a sandbox that can transcribe podcasts for you.

https://www.assemblyai.com/playground/source

That's pretty good, thanks!

Would you say the same about IRC that you do about Discord, had IRC its userbase size?

Yes. If an aughties open-source project had had no bug tracker, no mailing list, no Usenet group, but had written that "the team usually hangs out at #project at QuakeNet, feel free to drop in if you have any questions", I would have totally avoided using them if I could.

Software for separating voices (this means individual instrument types as well as human "voices") is pretty good. Its not good enough for performing over yet, but you can do karaoke at home with it. Its great for isolating or removing tracks for practice and transcription. I use Moises. https://moises.ai/ . When learning new music you can reduce or remove the voices you aren't trying to learn.

There's a free model released by OpenAI called Whisper that is excellent for transcribing voice to text in any context, not just music. Runs pretty well on modest hardware too to boot.

At any rate, I'm basking in the schadenfreude of being the Cassandra warning people about the incoming scourge of AI induced unemployment for at least 5 or 6 years now, only to be vindicated in real time.

You're all fucked, some just about to be fucked before the others.

You're all fucked, some just about to be fucked before the others.

Aren't you Indian?

Don't you have a ridiculous civil service which people kill to get into because once you're there, some quirk of iron rice bowls and pork barrels and constituency building has basically made Indian civil service jobs a sinecure where you never have to do any work but also you can't be fired, and this situation has persisted for 70 years despite the grinding poverty of all other sections of the Indian economy, because it's politically impossible to untangle this snarl?

And you think people are going to lose their jobs... because new labour saving tools become available?

Most jobs don't exist to fulfil tasks. Most jobs exist to fulfil government kayfabe. That an AI can perform a task is therefore completely irrelevant to the question of who has a job.

I've been pushing the meme of 'Government jobs are mostly middle class welfare' for years and people mostly just shrug and go about their day.

The Civil Service employs a tiny fraction of the population, and is incredibly hard to break into. It is by no means representative of the typical profession, which lacks said moats.

Idk about you, but my job is very much not bullshit, which at least helps me sleep better tonight except for when I'm lying sleepless knowing that that implies that some entity doing a better job at it than I can can actually replace me.

Also, there's a distinction between bullshit jobs that are de-facto jobs programs, and those that arise as a necessary consequence of poor regulation and incentives. The latter can be done by an AI no problem.

Also, there's a distinction between bullshit jobs that are de-facto jobs programs, and those that arise as a necessary consequence of poor regulation and incentives.

Do please elaborate how you make this distinction.

Digging a hole and filling it up again, pointless infrastructure projects such as in China = The former.

Title IX regulations (for a specific, albeit terrible reason, not intended simply as a jobs program), prompting the appointment of administrators to handle the demands = The latter.

I doubt either the government or the colleges wanted to simply create more jobs.

There is an economy outside of the bureaucracy, even in India. Furthermore, government patronage and kayfabe can be manipulated by using political power and wealth. If you have an AI that can get you power and wealth, you can then influence the government to do what you want.

Most jobs do exist to fill tasks. Some of them are surely bullshit, existing out of inertia or perhaps competition with a counterpart, and produce nothing. I think this is relatively rare. It’s much more common for a job to be marginally useful, generating value that would not have otherwise existed. Capitalism has a noise floor, but it’s not completely forgiving.

I'm just thinking of when I first publicly stated my thoughts. I've been aware of Humans Need Not Apply and similar notions for many years prior.

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