Beep boop - this is a robot. A new show has been posted to TWiT…
What are your thoughts about today’s show? We’d love to hear from you!
Pleasant show, as always!
That and (to avoid a but) the discussion about AI eradicating jobs (around 2:30 hrs) was, sorry to say, boomer deluxe:
- Benito makes a great point about learning skills and how automation kills jobs necessary to master valuable skills.
Responses:
- You don’t want to become good at what a computer can do.
- Equation of creative work to septic tank maintenance. (Which is a strange comparison for someone who’s son is hitting it out of the park big time with a TikTok career on sandwiches - but maybe I’ve missed something here.)
- Gutenberg-Moment about how it’s all a misconception and indeed there will be more jobs.
- Content is a commodity, we have to get past that. (Which is confusing, coming from a content creator who wants to start a movement for TikTok, cause creativity, freedom of speech, and voice and … ?)
Twist:
- Admission that it’s a real challenge.
- Cheeky bottom line “we’re too old - it won’t matter to us.”
I fully understand that this is an entertainment and news show and I know most things need to be taken with a grain of salt - heat of the moment, pressure to be entertaining and commit to a perspective and such. To be honest, I’m also half way on board with not demonizing this progress. Maybe I am also not getting half of the fine nuances of the discussion. Plus it’s really a complex topic and it’s also a bit mean to put spoken words on the scale after the discussion, buuut… well, let’s put it in this positive way: Benito is the man! Such an insightful and articulate person. It’s great to have him chime in more and more. His views are very valuable and often add perspectives that the main panel either misses or avoids. To have him only as an occasional voice from the off seems like only skirting his abilities, I think. He might be a great co-host!
That said: Jesus - another show going above 2:30 hrs - very generous. By God, please don’t listen to yourselves and stop producing content cause it’s commoditized and you need to get over it. Some people, many actually, like your content. And that it’s actually hand made. Because you have and create jobs in the field. Since they have not yet been replaced. Hurray!
2 Likes
I don’t think that was the intent. What was intended is making the point that some blue collar jobs (electrician, plumber, elevator repair person, and yes septic tank maintainers) have fallen out of favour, so there is higher demand for them, but they’re something that computers/AI won’t be doing in the near term. There is a bigger story here, it may be somewhat North America specific. See this video if you want to learn more
2 Likes
I’m not quite half way yet, so have that to come. But the discussion on reviews being written by AI… This is something an AI just can’t do. A review is to actually take something and use it, live with it and report on how it works.
How is an AI going to review a carpet? A car? (Well, if it is self-driving, maybe the AI can take it for a spin, or it can interview the car’s AI?) A set of tools? (How is the AI going to know, whether the tools are made of hardened metal, or will simply break apart, the first time you try and remove a screw from the wall? Or whether a torque wrench is accurate?)
We keep pushing AI at the wrong things, the “easy” things to do, even if the results are wildly inaccurate.
We’ve had AI in photography for years, it does a great job, and it is an area where AI can really help - even good photographers need to remove things from images, but not everyone is a good photographer and a Photoshop expert, so having an AI help with retouching an image is a great use, but it now falls under the radar, because AI doing something useful was yesterday, it is now boring, the exciting stuff is all in AI hallucinogenics!
Getting better AI recognition on a production line to filter out bad products? Something I first saw demonstrated in the early 80s, but still not quite there.
AI to help people with disabilities? Yeah, interesting, but niche, we need LLMs!?!?! LLMs are sexy!11!1!
We need LLMs to generate text about things they don’t understand! We need LLMs to create images nobody needs! Okay, that last one is questionable, Paul Thurrott makes good use of them for his blog, but it is still a niche. Most people don’t need to create images on a daily basis, but cleaning up photos of their kids or holiday shots etc. is an area where it could (and does) make a difference, but that is no longer considered AI, because it isn’t an LLM!
The application of AI to problems that really exist and could really benefit from AI seem to be being ignored, because the funding is all in LLMs and chatbots, not on actually solving every day problems.
Edit: Jeff said, when talking about the TikTok divestiture, that the fact the Chinese Government have stepped in and said that Byte Dance can’t sell the algorithm is proof that the Chinese Government is involved in Byte Dance…
Possibly, except… The US Government has often stepped in to stop the sale of companies or specific products, especially when going to the other side of the Iron Curtain. And, I think, if it was the other way round, the US Government would probably be bellyaching about the source code of Microsoft Windows being sold to a foreign entity, or the Facebook algorithm, or Google’s Search secret sauce having to be sold to a company in a different country…
1 Like
Edit: The callous way that the hosts replied to Benito’s accurate argument, now that I’ve listened to it, just show how far this podcast has shifted and after listening to this podcast for almost a year now I probably won’t be listening to it.
Jeff and Leo have went full tilt on AI and Jeff particularly just seems like a grump these days who doesn’t invite any alternative opinions about AI aside from his own.
I know people who are losing art jobs to AI as we speak and these people treat creatives the way Buzzfeed treated the news, as a some throw away utility. And the podcast itself is struggling in a sea of podcasts and yet they don’t seem to give a crap about artists and keep pushing AI as some boon.
I say it again, they need to have someone who is either skeptical of AI on these podcasts or actually sit down and listen to what artists say about AI and not some tech journalist.
The AI bias in these recent episodes can be scooped out of the air with a spoon. You have on one hand Leo who thinks AI should just consume every bit of information, ironic given his stance on privacy. On the other hand you have Jeff Jarvis who thinks guardrails on AI are a bad idea because (bad people will always do bad things which is logically the same thing as saying guns don’t kill people, people kill people ), and hasn’t ever offered a way that journalists are to make money when AI can summarize their work in seconds flat. A lot of it is just grand standing.
Influential people especially people who call themselves journalists should do more listening and reporting on these things rather than poo pooing any idea they don’t like because it doesn’t fit their preconceived notions of the world and immediately run to Gutenberg as if thats the end all be all of every argument against AI stealing creative work.
2 Likes
So how is a artist who lost concept art work supposed to feed themselves when a producer decides to "ideate the creative process " using midjourney ? Thats not a hypothetical, thats reality right now for many artists. Not to mention CG artists losing freelance work because game studios save a buck by using AI to make game assets, voice actors losing work because AI voice acting is cheaper (also not a hypothetical it’s something thats happening in the industry ), and I’ve even heard that science illustrators are slowly being replaced because people can make illustrations using AI.
1 Like
Oh, that’s easy, they just go out and clean all those septic tanks that @Leo was talking about… /s
I listened to the AI voices that were reading the stories on Audible. My first impression, less than 1 second in was, “this is artificial”. It was stilted, it was clearly computer generated and it lacked emotion. Going to the most boring man in the world, if you are doing a tutorial on something, you might get away with such a voice, but for a novel, or even a biography or history, I want a voice with feeling, that knows what they are talking about and can put emphasis in the right places, liven up the text and bring it to life.
1 Like
I’m no apologist for the “AI”. I understand it’s hard out there and that real people are facing real consequences. I was merely pointing out the harsh reality that in the end, you gotta go where the work is if you want to avoid unemployment because of AI induced redundancies. There is a longer term discussion to be had about things like basic personal income as a means to keep society afloat.
I think Paris is the voice against Jeff and Leo, but honestly it’s a pretty tedious discussion for everyone involved so I don’t blame her for not really wanting to talk about it every time it comes up.
Jeff has started many times that community > content, so I think that is what motivates his comments. Leo pretty much agrees with that assessment, I say on the forum he hosts as a club member. Paris doesn’t disagree from what I have heard in the past, although she remains skeptical that there is no place for copyright to play in these corporate interests. There, I think I have summed up the convos we missed!
Fwiw, I viewed the “we’re old and it won’t affect us” as a deflection to quickly transition away from the topic that is a downer and nobody has solutions for.
1 Like
Paris agrees more with Jeff really. The only time she’s called them out was through some light mocking which led to Jeff complaining that The information is behind a paywall (which is kind of a weird thing to say given that that’s what pays for Paris’s apartment and food.)
Which becomes even more deeply ironic when you realize that they really want club twit to succeed yet make snide remarks about people losing jobs to AI.
“Jeff has started many times that community > content” I’ll put this lightly, artists already have a community, community doesn’t pay the bills. Artists are producers and creators by the nature of their work, raw art even physical ones doesn’t sell the way it used to, so many work as animators, 3d artists, and concept artists. Those guys can’t just start a patreon even if they have large social media followers, since about less than 2 percent of people even care enough to pay for that art.
If Jeff really wanted to put his money for his mouth is he’d actually tell us what this great way working creative people can do to make money while facing diminishing job opportunities due to AI. Since he’s a no show we’ll have to assume he doesn’t have one and just likes to talk for sake of disagreeing.
“I viewed the “we’re old and it won’t affect us”” Nah, Jeff and Leo are eager to dispense their opinions on AI at the drop of a hat, the difference is that they believe in AI more than they seem to care about working creative people.
It reads to me more as, “we’ve secured our bag so this topic doesn’t affect us”. Look they’re entitled to their opinions, but they also have a opportunity here to actually be one of the few podcasts to actually go into real depth on the issue, they have a working reporter on their panel for crying out loud and a professor of journalism, yet they do so little listening most of the time. Look at the fact that none of them are even in this forum to clarify their comments. they don’t care, which is their right, but it’s a missed opportunity none the less.
Consider that the artists voices who are facing the brunt of job loss aren’t there to testify to the senate or get speaking fees for sharing their opinions at influential conferences. All they have is social media to vent their frustration as they see the entry level positions crumbling away, and younger artists are wondering how exactly they’re supposed to get their start in a shrinking industry, while senior artists are nervously looking at their phones realizing that they’ll be getting less and less calls soon. I mention artists, but this not even just artists who are suffering this.
Now while Jeff told Congress that AI could be beneficial in teaching people to read, as far as I can tell it hasn’t taught a single person how to read, instead it’s been used to flood Amazon with so many self published books that Amazon had to limit submission to 3 books a day. It’s been used to imitate working artists to the point that when you google someone like greg rutkowski many of the results aren’t even his artwork. It’s been used to create pornographjy of celebrities on a scale unseen before you could power a african country with the electricity burned by the gpus making all that porn.
For all the empirical promises of AI improving our water supply or accelerating technological progress, nothing of that sort has come about because talking to engineers or computer scientists they’re not impressed with AI’s abilities to actually help them do their jobs, no where the tech billionaires have emphasized their technology is by going after a human activity that doesn’t need to be automated.
1 Like
I think Jeff is, in part, a historian. He adores the Gutenberg Press after all. What he has said, as I heard him, is that over the decades, different technologies have been blamed as the one to end things the way they were at the time. He seems to think that AI is just another occurrence of this trend. In that sense, each old way of doing things is revitalized (or eliminated I suppose) by the people who were doing things the old way before they found whatever new way that works. This means he doesn’t know the way forward any more than anyone else, until after it happens and is also history.
Maybe to add a little bit perspective: the point discussed here was brought up after two and a half hours of lively and lighthearted discussion. I’d not be fully and thoroughly considerate and on point after such a long time performing either. Plus, if taken in context of the rest of the content, especially the share of agreeable and not agreeable content, I’d be hard pressed to point to ANY other media product that would be performing better that any of the TWiT podcasts. It’s impressive how listeners (me and many others here) can pick one five minute bit out of three hours and get hung up on it. It must be a testament to how much many of us love the shows. 
If we did not, we’d certainly not waste time discussing our thoughts here. Same for the show with Ed Citron as a guest. Certainly no hate, just a sign of how comfortable at least I got with the regular show family. Good to see him come back!
He’s also an active promoter of the technology. After all he has access to googles campus which invites him, so he can gawk at their latest AI tech. (Do artists get invites to googles campus to tell them what their AI tech is doing ? No) So his position is clear, he’s sided himself for whatever reason with companies who are selling a technology that by it’s nature is being promoted around Hollywood as not something to enhance human creativity, but to eliminate head count for the sake of extra profit. Instead of perhaps pointing this out people like Jeff repeatedly don’t do anything with their influential positions except sometimes play unpaid spokespeople (recall how much screen time Googles notebook LM got ? )
He’s not just interested in longtermism, he has actively wedged himself on the side of big tech and AI, see his testimony for proof. All of this while not a single time explaining how journalists, voice actors, concept artists are supposed to find work. I mean I don’t recall that his AI podcast has even interviewed a artist or anyone directly hit by AI.
You’re welcome to your opinions, but I find these framings quite uncharitable. Nobody has a real solution to these problems, and many in the show are skeptical of the role copyright can play. I’m glad they moved on quickly after introducing the topic.
You guys sound like modern day Luddites. “Smash the Jacquard looms - they’ll put weavers out of business.” Indeed they did. And while I have nothing but sympathy for the out of work weavers, you can’t deny that we have benefited from automated clothing manufacture.
The promise of AI is far greater. And, yes, there will be massive disruptions as a result. But that was true of the Internet, steam, and every other technological revolution.
I know it sounds terribly callous, but the alternative to technological progress is stagnation. And that’s not good either.
You can’t be surprised that a podcast network for and by technology enthusiasts would be in favor of technological progress.
1 Like
I don’t think it’s that. We’re here for technological progress and discussion of its effects.
Some of us were surprised of how little the impact on individual people - as raised by Benito - was taken seriously as problematic.
I think it’s a matter of misaligned context. Benito was mentioning specific cases and experiences while Jeff and Leo were in “big picture mode” - the mode in which the industry likes to “move fast and break things” and create “massive disruptions”. I find disruptions interesting and entertaining to observe when they impact big corporate incumbents. They can be devastating if they impact individuals. They can be devastating to communities if that one moonshot disruptor that would never pan out financially took out valuable infrastructure in its wake, before going bankrupt themselves. This warrants at least some addressing and sensibility.
In this week’s TWiT, I heard the idea to move some parts of policy making and political decisions to AI because the populace were not good enough at it. I’m sure it was said half jokingly. But if we hand over the creation of things to machines (done already in most cases), do the same for the creation of thoughts and creative works (which is happening currently), as well as the methodology to determine our destiny (which would be transitioning policy making to AI), we’d really have given up on ourselves.
Weaving yarn is one thing. Weaving thoughts and our future another. That’s why the Luddite analogy limps.
But - you know - after 2 1/2 hours of show, who can really take everything problematic what might benefit from taken as such? Fully sympathize. It’s a bit of a hair-in-the-soup discussion, in context of a lovely podcast.
2 Likes
It was interesting, that on TWiT this week, you moved around about 130° on this topic, finally echoing the argument I’ve been making for about 9 months, but that has been totally ignored.
I’m all for moving technology forward, and I actually find some of the things being released now, which assist people, as opposed to the generative AI itself. But the argument I have been making is, if you put all of the reporters out of business, who is going to make new, quality content for the AIs to learn from?
If nobody is covering issues in public, because all the companies that would pay for the content have gone, the AIs will be relying on content made by AIs and they will spiral into ever weirder reporting, as they echo off each other, because nobody is left to plug in real content.
That is something you actually touched on, but, unfortunately, you didn’t follow through on it, how this problem could be solved.
I have used various AIs LLMs, but I have yet to actually find a use for them in my day-to-day work or my leisure time.
I have been using Luminar Neo for a couple of years and the AI putting effects on images etc. is much more interesting and a real world example of how AI can help - I am a good photographer and I don’t need to touch up 90% of my images, but the remaining 10% need some help, but I am lousy when it comes to photo editing with tools like Photoshop - I can do some rudimentary changes, but anything complicated falls outside of my competence radius and Neo can help me.
That is a good use of AI. Making images from scratch doesn’t really interest me and, whilst some of the images created are impressive, it is the message, the feeling behind an image that is important and that is missing in AI generated art, plus the artifacts that AI can’t currently cope with, like the correct number of fingers or limbs being properly joined to a body, for example.
2 Likes
Luddites? I understand the mathematics behind this fairly well. (It isn’t a technology problem, it’s a math one. That’s why all the advances are posted on arxiv.) Please tell me what part of the probability distribution for tokenized words needs to be learned sufficiently to let your LLM govern societies.
You keep talking about AI when you usually mean LLMs. I find that discussion really frustrating. AI/ML (distinction only made depending on who signs the checks) has existed for a long time to answer questions about important problems for humankind.
I think just telling people they should find new jobs because you have a belief that their words will turn it into some super genius is bound to upset people. Calling them Luddites for not sharing your belief is rude. Peace.
2 Likes
Luddites weren’t against technological progress, they were against the unjust system that used the looms to benefit a few without sharing out the profits. If you think AI is great, whatever that’s your opinion, but sticking your head in the sand and saying "Yeah whatever people always go out of work " is as silly as waving your hands and saying "Pah, climate change, the climate is always changing ".
Kudos to you for knowing it sounds "callous ", but your opinion would be far less privileged if you were on the other side of the line facing job loss because your boss decided to train a AI model to do your job.
Also the challenge stands to you and Jeff to give creative people a reply as to how they’re supposed to make money as they face less work
Leo seems to not really understand the difference between LLMs and other forms of AI like the ones used to predict protein folding.
To give an example, I know of a AI that uses a GAN to look at snap shots of crystal structures and makes predictions about the structure of opaque crystals. This AI uses exactly 0 reddit comments in it’s training data, instead it needs domain specific data.
So this idea that Chatgpt and Sora.ai are somehow useful to humanity is laughable. In what way is a leisure creative activity of humans like art improved when a machine can automate it ? The only people eager to deploy it are producers and studio execs and all kinds of middle level graphics companies who want to increase profit margins.
So his response shows that he doesn’t really fully seem to understand what AI is and what it’s used for and hasn’t made spent time considering how automating creative work is beneficial. His privilege is showing.
2 Likes