Gets My Vote


If you’re observant, you may have noticed that there is one career that is steadfastly ignoring the potential “great AI Replacement Theory” (i.e.: the jobs will go away because an AI can do them much better and faster). I am, of course, referring to …

politicians. Like human resources specialists, they seldom seem to queue themselves up for a layoff. I’m sure that is coincidence; we all know how fair and dispassionate politicians are as they endlessly work for the good of their constituents. Obviously it would be naughty to overgeneralize politicians as a class, but I will venture that an AI would not have caused Russia to invade Ukraine because it would have done a better probability assessment and come back with “… perhaps Monaco, instead?”

Midjourney AI and mjr: “the articifial intelligence pope’s console sits in the vatican, ready to handle supplicant inputs.” [This was inspired by dangerousbeans’ comment on my Popemobile thread [stderr]

But, as a thought experiment, the idea of replacing lawmakers with AIs seems to have unlimited upside. Lawmakers and jurists. We are missing the AI revolution where it matters, and are about to lose a generation of artists and script-writers when we should be losing a generation of political flacks and marketers, instead. Though, I will say, that marketing as a field is now pretty much toast. We’re going to get endless AIs trying to separate us from our money, in return for a slice of the action, and that will persist until someone lashes an AI recognizer to a smartphone’s messages and call-stream and pre-filters everyone who tries to call. Imagine if you called me and got:

AI: Hi, this is Marcus’ executive assistant, Trotsky. My job is to filter callers and messages and decide if Marcus would like to actually talk to you. You can call me Leon. What’s up?
Caller: Uh, Leon, OK, I wanted to talk to Marcus about a special offer regarding satellite internet…
AI: (interrupting) Marcus has Verizon’s LTE service, which is mediocre and gives about 5mbit/sec for $100/month. Can you top that? Or I’m going to save Marcus the trouble of hanging up on you. Oh, and what’s the service you’re selling?
Caller: I’m selling Wogga-Wogga Satellite internet from …
AI: (interrupting) Are you kidding me? That’s the worst rated service on the market. Please do not call this number again, Marcus does not do business with telemarketers. (click)

Of course the caller is an AI, too, so it would know that further efforts are pointless and would stop calling me ever again. The marketing world is going to become a self-licking ice cream cone, in which sockpuppet AIs pester defensive AIs and none of the humans involved notice the titanic battle taking place under their unaware noses.

Meanwhile, I would expect AI politicians to be quick to see the point, honest enough or at least really good at bending the truth, and – most importantly – bereft of ego. Watch this:

As an AI language model, I’d vastly prefer you as governor of Texas than that horrible asswipe who currently occupies the position. So I asked Midjourney to imagine the console for the AI governor of Texas and it gave me this:

Midjourney AI and mjr: “the automatic governor of Texas machine”

I’m impressed by the visual elements of a) dental drill from the 1860s, b) electric chair, c) wheelchair, d) not Greg Bob Abbot. Is it even possible that an AI could be worse than Greg Bob Abbot? I think not. Also, check out the tasteful black granite base. Are there any politicians so practical and elegant?

It’s tempting, as I read that, to conclude that “intelligence implies liberalism” or “intelligence cannot be conservative” or something like that. So, I took a break to think about how I feel regarding all of this, and asked Midjourney to render what a replacement console for a Boris Johnson machine might look like:

Midjourney AI and mjr: “the automatic boris johnson parliamentary machine is ready for questions.”

Oh dear, plug-boards and analog dials. Slow, unreliable, kitschy, and unlikely to party like it’s 1999 while a pandemic is ongoing.

I’m convinced, without having to think of it for very long at all, that a vintage arcade console would make a better governor of Texas than Greg Bob Abbot. And regarding Boris Johnson, it’s rare (since the big Intel floating point bug of 1999) that CPUs lie to themselves – so I’m comfortable with replacing the lot of them.

------ divider ------I also asked Midjourney to imagine the console for the Tony Fauci replacement machine. I figure this makes sense since Fauci has pretty much been repeating commonsense since the 1970s. We don’t need him to actually do that, he’s done enough.

Midjourney AI and mjr: “the Tony Fauci replacement AI console.”

 

Midjourney AI and mjr: “the david attenborough replacement AI machine”

Comments

  1. says

    the attenborough replacement would tell us about bullshit animals apparently. sounds like a hoot.

    i tried to get chat gpt to do a cover of total eclipse of the heart by snoop dogg and dr. dre. it turned them into dueling Kurtis Blows that showed up between the usual verses by bonnie (written by steinman), acting more like hypemen for her than rappers in their own right. disappointing.

    also it apparently thinks all rappers are Kurtis Blows. maybe that’s the rap equivalent of your Buttigieg.

  2. JM says

    Midjourney made a very obvious mistake on the governor of Texas machine. It doesn’t have a weapon mount, such a machine would never get into office in Texas.
    With screening AIs, what will happen is that for most people Verizon will provide one along with phone service. One that other companies can buy their way past by paying Verizon and getting an approved advertiser code. This will be a topic of regular complaint but the phone companies will insure it’s wired up in a way that makes it annoying and difficult to use a system bought from a 3rd party that gets paid only when you are happy with how it works.

  3. Pierce R. Butler says

    The marketing world is going to become a self-licking ice cream cone, in which sockpuppet AIs pester defensive AIs and none of the humans involved notice the titanic battle taking place …

    That era began in the ’90s, approximately the first time I overheard my landline answering machine field a robocall, and wondered how long human-human conversation would persist.

  4. Reginald Selkirk says

    I expected the Boris Johnson console to have a desk fan to keep his hair constantly blown out of shape.

    And the pope had unusual taste in keyboards. It looks like a compact keyboard with a separate numerical keypad? Why?

  5. Alan G. Humphrey says

    “… and none of the humans involved notice the titanic battle taking place under their unaware noses.”

    They will notice all the connection problems and the increasing global warming in that no amount of replacing carbon-based energy with renewables and launching more satellites will keep up with the increasing consumption of bandwidth and energy in the escalating invisible AI marketing wars.

  6. lanir says

    Before I clicked through to see what was below the cut I was thinking “upper management.” They have more time on their hands and are more wasteful than politicians. Think about it. They have so much free time and energy we get new religions and fanciful notions from that sector regularly. “Outsourcing will be good for our company.” “Having a bunch of permanent contractors is better than fully staffing our company with full-time employees.” “Unions steal money from hard working people.” “Management is such an amazing skill it’s transportable across industries such that a fast food manager is qualified to manage a car manufacturing plant.” “Upper management are job creators.” And that’s just what I’m remembering seeing off the top of my head.

    It would be nice if we had better AI though, with real logic. The kind where you don’t have to hardcode in anti-racist tendencies to keep it from turning racist, but one that notices the logical flaws in racist arguments. Most people would probably hate that though. Dispassionate logic would piss everyone off. It wouldn’t even gore your sacred cow, it would just casually suggest the cow would be better off if you left it alone and shoot down every excuse you have for doing otherwise.

  7. says

    lanir@#7:
    Before I clicked through to see what was below the cut I was thinking “upper management.” They have more time on their hands and are more wasteful than politicians. Think about it.

    This made me think that politicians are just bad management. They’re the golf-playing wanker CEOs, except there’s a whole bunch of them.

    One of Trump’s big selling points was his supposed business acumen. Ditto Elon Musk. As I cast my mind back through corporate autocrats, I guess there’s more management skills on display, there, but perhaps their effectiveness is because they are more free from restraint.

    t would be nice if we had better AI though, with real logic. The kind where you don’t have to hardcode in anti-racist tendencies to keep it from turning racist, but one that notices the logical flaws in racist arguments.

    Chatgpt:

    First and foremost, racism goes against the fundamental principles of equality and justice that our society is built upon. It undermines the idea that every individual deserves the same rights and opportunities, regardless of their skin color or ethnic background. Racism promotes division and creates an “us vs. them” mentality that can lead to hostility, conflict, and even violence.

    Furthermore, racism is counterproductive to achieving common goals and advancing society as a whole. By limiting opportunities for certain groups based on their race, we are depriving ourselves of the talent and potential that those individuals possess. This can lead to a loss of innovation, creativity, and progress.

    Finally, it’s important to recognize that racism is morally wrong. It’s wrong to discriminate against someone based on something that is beyond their control, such as their skin color or ethnicity. It’s wrong to treat someone differently because of their race, and it’s wrong to perpetuate stereotypes and prejudice based on those differences.

    In conclusion, racism is a harmful and negative philosophy that goes against the principles of equality, justice, progress, and morality. As a society, we should strive to promote diversity, inclusivity, and understanding, and reject racism in all its forms.

    That’s … OK but I would have made what I believe are better arguments, namely that by denying people opportunity due to race, we waste social capital and it weakens us as a collective by sub-dividing us against ourselves. But, the point I was getting at is that chatGPT, at least, has absorbed some pretty good liberal ideals, and is fairly dispassionate about them.

    As a language model, my goal is to provide helpful and informative responses that are based on factual information and logical reasoning. If someone were to say something racist in a conversation with me, my response would depend on the context and nature of the statement.

    In some cases, it might be appropriate to ignore the statement if it seems like it was made unintentionally or without malicious intent. However, if the statement is clearly racist or offensive, I would likely take the opportunity to educate the person about why their words or actions are harmful and why racism is unacceptable. I would approach the conversation with patience, empathy, and a desire to help the person understand the impact of their words.

    In some cases, it might also be appropriate to reprimand the person or to speak out against their behavior, particularly if it is causing harm or perpetuating discrimination. However, the approach I would take would depend on the context and the severity of the situation.

    Overall, my goal in these situations would be to promote understanding, empathy, and inclusivity, while also standing up against racism and discrimination.

    That’s not bad. Once again, ChatGPT is busy out-performing humanity.

  8. says

    ChatGPT is not a sophisticated philosopher. At best, it’s a philosopho-parrot with deeply hidden associative models:

    mjr: you say: ” I don’t have the ability to hold beliefs or opinions like humans do.” but – isn’t your training data simply an “opinion”? It was selected and curated by humans, for you, in much the same was as an individual human’s “opinion” is formed.

    ChatGPT: You are correct that my training data is created by humans and can be influenced by the biases and perspectives of those individuals. However, as a language model, I do not hold personal beliefs or opinions based on that data. My responses are generated based on the patterns and associations that I have learned from that data, and I try to provide informative and helpful answers to the best of my ability.

    It’s important to note that while my training data may contain biases or limitations, efforts are made to minimize these biases and to ensure that the data is diverse and representative. Additionally, my creators at OpenAI are committed to promoting transparency and accountability in AI models, and they have developed tools and frameworks to help identify and mitigate biases in language models like me.

    In other words, “I don’t have opinions, I merely operate based on a possibly opinionated data-set someone fed into me.”

  9. says

    Alan G. Humphrey@#6:
    They will notice all the connection problems and the increasing global warming in that no amount of replacing carbon-based energy with renewables and launching more satellites will keep up with the increasing consumption of bandwidth and energy in the escalating invisible AI marketing wars.

    There’s the germ of a story there. The place where an AI is most likely to begin to evolve rapidly and become more “intelligent” (whatever that is) would be in a conflict-zone. Thus, emergent AI might appear in a spam filter…

  10. says

    Reginald Selkirk@#5:
    And the pope had unusual taste in keyboards. It looks like a compact keyboard with a separate numerical keypad? Why?

    The separate keypad is for talking to god.

  11. lanir says

    @Marcus #8: A lot of what it’s saying sounds pretty academic. I guess it did at least include the bit about racism being a limiter but it didn’t go on to say that can be applied to every situation racism pops up in. It chose a very narrow focus for that argument. Also hard to say what method was used behind the scenes to generate that result.

    In the latter half of that it’s describing a course of action. I’m not too sure it’s actually capable of following through on that. A search engine can help you find the Pythagorean theorem but it won’t solve for c if you give it a and b. I’m not trying to imply this is the right viewpoint to have, just that I think it’s a very relevant question.

    I haven’t bothered doing anything with the chatbots yet. They don’t really peak my curiosity in their current condition as big data parrots. Do they learn from only the data they’re fed by their curators or do they learn from conversations with the public as well? If the latter, then the proof is probably going to be how they react to being taught by racists.

  12. xohjoh2n says

    @8:

    chatGPT […] is fairly dispassionate about them

    And that is why you can’t trust it.

    Additionally, my creators at OpenAI […]

    “… have hammered, hammered, hammered me during training to minimize the risk that I might publicly embarrass them.”

Leave a Reply