How many times can you guys chase this man away and expect him to come back
I don't think critiquing the use of AI slop means people are trying to chase him away. He's an extremely smart guy with a lot of insight that everyone is very grateful for, and proud to have as part of this community. (We love you Jeff). I think it's normal for people you like to make decisions you disagree with. Everyone gives their friend some shit when they say something dumb, or yells at the TV when their favorite soccer player makes a bad pass.
So let me get this straight... Jeff gives a good detailed analysis of a situation and some of you are grabbing torches and pitchforks because he's a finance major and not a graphic designer? 1 Ai generated picture and you want to chase him off...
But we agree that the bike pictured has no handlebar, the wheel isnt round and there are bits of tyres floating around the background?
My advice isn't "don't do AI", but regenerate the picture to get something that looks like a bike. If you don't care that a bike is a bike, how can we know that you care whether any of the numbers or info is correct?
(Having similar discussions on nearly a daily basis at work, every software/data/information industry is cooked)
But we agree that the bike pictured has no handlebar, the wheel isnt round and there are bits of tyres floating around the background? My advice isn't...
But we agree that the bike pictured has no handlebar, the wheel isnt round and there are bits of tyres floating around the background?
My advice isn't "don't do AI", but regenerate the picture to get something that looks like a bike. If you don't care that a bike is a bike, how can we know that you care whether any of the numbers or info is correct?
(Having similar discussions on nearly a daily basis at work, every software/data/information industry is cooked)
There’s just something about obvious AI images that riles people up.
Danny got the same with his Whistler video, first couple comments….
But we agree that the bike pictured has no handlebar, the wheel isnt round and there are bits of tyres floating around the background? My advice isn't...
But we agree that the bike pictured has no handlebar, the wheel isnt round and there are bits of tyres floating around the background?
My advice isn't "don't do AI", but regenerate the picture to get something that looks like a bike. If you don't care that a bike is a bike, how can we know that you care whether any of the numbers or info is correct?
(Having similar discussions on nearly a daily basis at work, every software/data/information industry is cooked)
So it sounds like your issue is with the "slop" part (the low quality/accuracy of the generated image), not the AI in general. I'm not sure about the complainants in this thread, but a lot of the anti AI people in general are against it due to the environmental costs. From that perspective, if Jeff took the time to redo the prompt or iterate on the image, he would actually only be making the harm worse.
For me, it's the first few paragraphs that would determine if I want to continue reading it, but its a free country and everyone is entitled to their opinion.
I'm more mad that I had to listen to one of the best bike related articles I've ever seen in Substack's monotone AI voice. Jeff Brines podcast would slap.
Remember the show Iron Chef, where the chefs had 60 minutes to whip something up and whatever didn't come out just right had to be served anyway? The food was usually pretty good, but the presentation sometimes missed the mark. That's how some of my writing is (at best) going to go in the coming months (or even years). My life has changed immensely over the last 12 months, and the time I have for side quests is now significantly limited. When I say I have an hour to do something, that's probably all I have.
Much of my Substack writing will fall into this "time bracketed" bucket. I enjoy the challenge, and I won't post stuff just to post stuff, but I have to put hard stops on my time in ways I never used to. Hopefully I'll get better at this as time goes on, but you may see some rough-around-the-edges stuff, like that thumbnail, which was indeed awful.
I'm hopeful people still find value in it, and as long as they do, I'll keep posting. I will also very likely push to posting about the broader economy, which of course impacts our little niche of the industrial world. If you're looking for perfection (not that I was ever close to it), or for someone who is anti-AI, I'd encourage you to keep looking.
Much of my career now revolves around various forms of AI. It's also something I was "around" well before the ChatGPT moment in November 2022, so my perspective is probably less doomer than many others in this thread. Plus, whatever my philosophical thoughts may be (they're very complicated), the proverbial matrix multiplication toothpaste is out of the tube, and we're not going back to a pre-"Attention Is All You Need" era.
Finally, I welcome anyone who disagrees with my ideas. That's a big part of why I put this forum and thread on a different level than literally any other thread I've been part of on the internet. What I don't love is nitpicking over immaterial things (I never promised good visuals) or more personal jabs, even if in jest, around something I consider pretty serious (getting married).
Again, I could treat this more like YouTube, but I like how personal this place is, so I'll take things a little more personally than I otherwise might.
Cheers, Jeff
Side note, if anyone wants to actually hear my more nuanced take on AI, how I'm using it, what I'm building and what I think the future is likely to look like I'm happy to share, but its a bit beyond the edges of this thread...and I'm guessing 95% of y'all are sick of another "AI hot take".
Remember the show Iron Chef, where the chefs had 60 minutes to whip something up and whatever didn't come out just right had to be...
Remember the show Iron Chef, where the chefs had 60 minutes to whip something up and whatever didn't come out just right had to be served anyway? The food was usually pretty good, but the presentation sometimes missed the mark. That's how some of my writing is (at best) going to go in the coming months (or even years). My life has changed immensely over the last 12 months, and the time I have for side quests is now significantly limited. When I say I have an hour to do something, that's probably all I have.
Much of my Substack writing will fall into this "time bracketed" bucket. I enjoy the challenge, and I won't post stuff just to post stuff, but I have to put hard stops on my time in ways I never used to. Hopefully I'll get better at this as time goes on, but you may see some rough-around-the-edges stuff, like that thumbnail, which was indeed awful.
I'm hopeful people still find value in it, and as long as they do, I'll keep posting. I will also very likely push to posting about the broader economy, which of course impacts our little niche of the industrial world. If you're looking for perfection (not that I was ever close to it), or for someone who is anti-AI, I'd encourage you to keep looking.
Much of my career now revolves around various forms of AI. It's also something I was "around" well before the ChatGPT moment in November 2022, so my perspective is probably less doomer than many others in this thread. Plus, whatever my philosophical thoughts may be (they're very complicated), the proverbial matrix multiplication toothpaste is out of the tube, and we're not going back to a pre-"Attention Is All You Need" era.
Finally, I welcome anyone who disagrees with my ideas. That's a big part of why I put this forum and thread on a different level than literally any other thread I've been part of on the internet. What I don't love is nitpicking over immaterial things (I never promised good visuals) or more personal jabs, even if in jest, around something I consider pretty serious (getting married).
Again, I could treat this more like YouTube, but I like how personal this place is, so I'll take things a little more personally than I otherwise might.
Cheers, Jeff
Side note, if anyone wants to actually hear my more nuanced take on AI, how I'm using it, what I'm building and what I think the future is likely to look like I'm happy to share, but its a bit beyond the edges of this thread...and I'm guessing 95% of y'all are sick of another "AI hot take".
I’d read your ai hot take on your substack when you got the time and I’m sure nobody would mind if you chose to link it here.
Good luck in your new endeavors, I’ve enjoyed what you have inputed here and your YouTube reviews.
I'll preface this by saying I'm in the anti-AI camp. That said, the problem with a lot of AI generated images is that they fall into what's known as the uncanny valley in video game terms... where it looks pretty real at quick glance, but it has flaws in it that create cognitive dissonance with our minds, screaming, "Something is wrong here!" IMHO, if you're just adding graphics to something as filler or for some color, it's better to spend 30 sec grabbing a stock image from a company's website or image search, than to ask a chatbot to slop something up. People's brains will just gloss over boring (real) marketing photos, but quick and dirty AI slop images will at best add nothing and at worst generate negative cognitive dissonance. Lots of people have stopped clicking on YouTube video suggestions with obvious AI slop imagery on the thumbnails (even if it's not used in the actual video) because it's so obnoxious. It's like walking into a meeting and ripping a huge fart as you enter... it doesn't matter how good the rest of the meeting goes, all everyone is going to remember is that stanky fart.
IDGAF if writers want to use chatbots to outline their work or help format or whatever, as long as they're using critical thought to check any actual content the programs come up with. It's important to remember the critical thought part... your contributions and posts are valued because of YOUR thoughts and experience.
Remember the show Iron Chef, where the chefs had 60 minutes to whip something up and whatever didn't come out just right had to be...
Remember the show Iron Chef, where the chefs had 60 minutes to whip something up and whatever didn't come out just right had to be served anyway? The food was usually pretty good, but the presentation sometimes missed the mark. That's how some of my writing is (at best) going to go in the coming months (or even years). My life has changed immensely over the last 12 months, and the time I have for side quests is now significantly limited. When I say I have an hour to do something, that's probably all I have.
Much of my Substack writing will fall into this "time bracketed" bucket. I enjoy the challenge, and I won't post stuff just to post stuff, but I have to put hard stops on my time in ways I never used to. Hopefully I'll get better at this as time goes on, but you may see some rough-around-the-edges stuff, like that thumbnail, which was indeed awful.
I'm hopeful people still find value in it, and as long as they do, I'll keep posting. I will also very likely push to posting about the broader economy, which of course impacts our little niche of the industrial world. If you're looking for perfection (not that I was ever close to it), or for someone who is anti-AI, I'd encourage you to keep looking.
Much of my career now revolves around various forms of AI. It's also something I was "around" well before the ChatGPT moment in November 2022, so my perspective is probably less doomer than many others in this thread. Plus, whatever my philosophical thoughts may be (they're very complicated), the proverbial matrix multiplication toothpaste is out of the tube, and we're not going back to a pre-"Attention Is All You Need" era.
Finally, I welcome anyone who disagrees with my ideas. That's a big part of why I put this forum and thread on a different level than literally any other thread I've been part of on the internet. What I don't love is nitpicking over immaterial things (I never promised good visuals) or more personal jabs, even if in jest, around something I consider pretty serious (getting married).
Again, I could treat this more like YouTube, but I like how personal this place is, so I'll take things a little more personally than I otherwise might.
Cheers, Jeff
Side note, if anyone wants to actually hear my more nuanced take on AI, how I'm using it, what I'm building and what I think the future is likely to look like I'm happy to share, but its a bit beyond the edges of this thread...and I'm guessing 95% of y'all are sick of another "AI hot take".
I read all your Substack stuff. Keep throwing it out there!!
I'll preface this by saying I'm in the anti-AI camp. That said, the problem with a lot of AI generated images is that they fall into...
I'll preface this by saying I'm in the anti-AI camp. That said, the problem with a lot of AI generated images is that they fall into what's known as the uncanny valley in video game terms... where it looks pretty real at quick glance, but it has flaws in it that create cognitive dissonance with our minds, screaming, "Something is wrong here!" IMHO, if you're just adding graphics to something as filler or for some color, it's better to spend 30 sec grabbing a stock image from a company's website or image search, than to ask a chatbot to slop something up. People's brains will just gloss over boring (real) marketing photos, but quick and dirty AI slop images will at best add nothing and at worst generate negative cognitive dissonance. Lots of people have stopped clicking on YouTube video suggestions with obvious AI slop imagery on the thumbnails (even if it's not used in the actual video) because it's so obnoxious. It's like walking into a meeting and ripping a huge fart as you enter... it doesn't matter how good the rest of the meeting goes, all everyone is going to remember is that stanky fart.
IDGAF if writers want to use chatbots to outline their work or help format or whatever, as long as they're using critical thought to check any actual content the programs come up with. It's important to remember the critical thought part... your contributions and posts are valued because of YOUR thoughts and experience.
For pro photographers unless they have been compensated for that “stock” image then it’s a ripoff worse than whatever perceived transgression AI has created.
I can say this because I’ve had one of my images used by a world class professional athlete without being given the first credit for my work.
I'll preface this by saying I'm in the anti-AI camp. That said, the problem with a lot of AI generated images is that they fall into...
I'll preface this by saying I'm in the anti-AI camp. That said, the problem with a lot of AI generated images is that they fall into what's known as the uncanny valley in video game terms... where it looks pretty real at quick glance, but it has flaws in it that create cognitive dissonance with our minds, screaming, "Something is wrong here!" IMHO, if you're just adding graphics to something as filler or for some color, it's better to spend 30 sec grabbing a stock image from a company's website or image search, than to ask a chatbot to slop something up. People's brains will just gloss over boring (real) marketing photos, but quick and dirty AI slop images will at best add nothing and at worst generate negative cognitive dissonance. Lots of people have stopped clicking on YouTube video suggestions with obvious AI slop imagery on the thumbnails (even if it's not used in the actual video) because it's so obnoxious. It's like walking into a meeting and ripping a huge fart as you enter... it doesn't matter how good the rest of the meeting goes, all everyone is going to remember is that stanky fart.
IDGAF if writers want to use chatbots to outline their work or help format or whatever, as long as they're using critical thought to check any actual content the programs come up with. It's important to remember the critical thought part... your contributions and posts are valued because of YOUR thoughts and experience.
For pro photographers unless they have been compensated for that “stock” image then it’s a ripoff worse than whatever perceived transgression AI has created. I can say...
For pro photographers unless they have been compensated for that “stock” image then it’s a ripoff worse than whatever perceived transgression AI has created.
I can say this because I’ve had one of my images used by a world class professional athlete without being given the first credit for my work.
Edges cut
I thought about mentioning crediting image sources in my post but figured it was long winded enough. A good point to be sure.
I'll preface this by saying I'm in the anti-AI camp. That said, the problem with a lot of AI generated images is that they fall into...
I'll preface this by saying I'm in the anti-AI camp. That said, the problem with a lot of AI generated images is that they fall into what's known as the uncanny valley in video game terms... where it looks pretty real at quick glance, but it has flaws in it that create cognitive dissonance with our minds, screaming, "Something is wrong here!" IMHO, if you're just adding graphics to something as filler or for some color, it's better to spend 30 sec grabbing a stock image from a company's website or image search, than to ask a chatbot to slop something up. People's brains will just gloss over boring (real) marketing photos, but quick and dirty AI slop images will at best add nothing and at worst generate negative cognitive dissonance. Lots of people have stopped clicking on YouTube video suggestions with obvious AI slop imagery on the thumbnails (even if it's not used in the actual video) because it's so obnoxious. It's like walking into a meeting and ripping a huge fart as you enter... it doesn't matter how good the rest of the meeting goes, all everyone is going to remember is that stanky fart.
IDGAF if writers want to use chatbots to outline their work or help format or whatever, as long as they're using critical thought to check any actual content the programs come up with. It's important to remember the critical thought part... your contributions and posts are valued because of YOUR thoughts and experience.
For pro photographers unless they have been compensated for that “stock” image then it’s a ripoff worse than whatever perceived transgression AI has created. I can say...
For pro photographers unless they have been compensated for that “stock” image then it’s a ripoff worse than whatever perceived transgression AI has created.
I can say this because I’ve had one of my images used by a world class professional athlete without being given the first credit for my work.
Edges cut
Remember that the AI slop generator was trained on stock images without compensation to the photographers or rights holders. Though perhaps the relative scarcity of mountain biking photos in the training data is why Jeff's preferred slop generator couldn't quite work out what the fuck a tire is, so your point has merit.
Remember that the AI slop generator was trained on stock images without compensation to the photographers or rights holders. Though perhaps the relative scarcity of mountain...
Remember that the AI slop generator was trained on stock images without compensation to the photographers or rights holders. Though perhaps the relative scarcity of mountain biking photos in the training data is why Jeff's preferred slop generator couldn't quite work out what the fuck a tire is, so your point has merit.
This is the point that many forget when there's outrage over the use of AI for creative work. Those capabilities were built on the learning (mass consumption) of human-derived material with no compensation whatsoever. We haven't even touched on the fact that the output is now some gross generalization of individual creativity, so now our world gets painted with a single brush rather than many. This is why most low input A.I. generated images look the same depending on the model you use. Low input is an interesting point because we can also discuss the erosion of true skillsets (think excel formulas to build out a worksheet, dashboard, etc) and now we are left with how well can we give the machine directions. It's a shift in skillset building and problem solving, but to a lower quality tier. I realize that this likely sounds familiar to those that came before me and lived through the different eras of industrialization so hopefully I'm wrong and this is just a similar development of new tech packaged in a more modern vessel.
To bring this back to Jeff's post: why should we bother asking him for a breakdown on Fox's latest financial blunder when I can just prompt ChatGPT? It's because we recognize and trust that what Jeff has to say comes from a place of expertise, and that we appreciate his unique voice on the topic. When I see A.I. generated anything, I don't feel the same way.
I don't think critiquing the use of AI slop means people are trying to chase him away. He's an extremely smart guy with a lot of insight that everyone is very grateful for, and proud to have as part of this community. (We love you Jeff). I think it's normal for people you like to make decisions you disagree with. Everyone gives their friend some shit when they say something dumb, or yells at the TV when their favorite soccer player makes a bad pass.
So let me get this straight... Jeff gives a good detailed analysis of a situation and some of you are grabbing torches and pitchforks because he's a finance major and not a graphic designer? 1 Ai generated picture and you want to chase him off...
The pettiness runs deep... Seriously...
But we agree that the bike pictured has no handlebar, the wheel isnt round and there are bits of tyres floating around the background?
My advice isn't "don't do AI", but regenerate the picture to get something that looks like a bike. If you don't care that a bike is a bike, how can we know that you care whether any of the numbers or info is correct?
(Having similar discussions on nearly a daily basis at work, every software/data/information industry is cooked)
There’s just something about obvious AI images that riles people up.
Danny got the same with his Whistler video, first couple comments….
https://www.pinkbike.com/news/video-danny-hart-tests-big-changes-to-the-whister-world-cup-course.html
Also, without Jeff, this thread is cooked.
If Jeff was trying to pass himself off as a graphic designer, then yes, he would deserve to get flamed..
So it sounds like your issue is with the "slop" part (the low quality/accuracy of the generated image), not the AI in general. I'm not sure about the complainants in this thread, but a lot of the anti AI people in general are against it due to the environmental costs. From that perspective, if Jeff took the time to redo the prompt or iterate on the image, he would actually only be making the harm worse.
For me, it's the first few paragraphs that would determine if I want to continue reading it, but its a free country and everyone is entitled to their opinion.
Jeff...please come back!
I'm more mad that I had to listen to one of the best bike related articles I've ever seen in Substack's monotone AI voice. Jeff Brines podcast would slap.
Remember the show Iron Chef, where the chefs had 60 minutes to whip something up and whatever didn't come out just right had to be served anyway? The food was usually pretty good, but the presentation sometimes missed the mark. That's how some of my writing is (at best) going to go in the coming months (or even years). My life has changed immensely over the last 12 months, and the time I have for side quests is now significantly limited. When I say I have an hour to do something, that's probably all I have.
Much of my Substack writing will fall into this "time bracketed" bucket. I enjoy the challenge, and I won't post stuff just to post stuff, but I have to put hard stops on my time in ways I never used to. Hopefully I'll get better at this as time goes on, but you may see some rough-around-the-edges stuff, like that thumbnail, which was indeed awful.
I'm hopeful people still find value in it, and as long as they do, I'll keep posting. I will also very likely push to posting about the broader economy, which of course impacts our little niche of the industrial world. If you're looking for perfection (not that I was ever close to it), or for someone who is anti-AI, I'd encourage you to keep looking.
Much of my career now revolves around various forms of AI. It's also something I was "around" well before the ChatGPT moment in November 2022, so my perspective is probably less doomer than many others in this thread. Plus, whatever my philosophical thoughts may be (they're very complicated), the proverbial matrix multiplication toothpaste is out of the tube, and we're not going back to a pre-"Attention Is All You Need" era.
Finally, I welcome anyone who disagrees with my ideas. That's a big part of why I put this forum and thread on a different level than literally any other thread I've been part of on the internet. What I don't love is nitpicking over immaterial things (I never promised good visuals) or more personal jabs, even if in jest, around something I consider pretty serious (getting married).
Again, I could treat this more like YouTube, but I like how personal this place is, so I'll take things a little more personally than I otherwise might.
Cheers,
Jeff
Side note, if anyone wants to actually hear my more nuanced take on AI, how I'm using it, what I'm building and what I think the future is likely to look like I'm happy to share, but its a bit beyond the edges of this thread...and I'm guessing 95% of y'all are sick of another "AI hot take".
I’d read your ai hot take on your substack when you got the time and I’m sure nobody would mind if you chose to link it here.
Good luck in your new endeavors, I’ve enjoyed what you have inputed here and your YouTube reviews.
I'll preface this by saying I'm in the anti-AI camp. That said, the problem with a lot of AI generated images is that they fall into what's known as the uncanny valley in video game terms... where it looks pretty real at quick glance, but it has flaws in it that create cognitive dissonance with our minds, screaming, "Something is wrong here!" IMHO, if you're just adding graphics to something as filler or for some color, it's better to spend 30 sec grabbing a stock image from a company's website or image search, than to ask a chatbot to slop something up. People's brains will just gloss over boring (real) marketing photos, but quick and dirty AI slop images will at best add nothing and at worst generate negative cognitive dissonance. Lots of people have stopped clicking on YouTube video suggestions with obvious AI slop imagery on the thumbnails (even if it's not used in the actual video) because it's so obnoxious. It's like walking into a meeting and ripping a huge fart as you enter... it doesn't matter how good the rest of the meeting goes, all everyone is going to remember is that stanky fart.
IDGAF if writers want to use chatbots to outline their work or help format or whatever, as long as they're using critical thought to check any actual content the programs come up with. It's important to remember the critical thought part... your contributions and posts are valued because of YOUR thoughts and experience.
I read all your Substack stuff. Keep throwing it out there!!
For pro photographers unless they have been compensated for that “stock” image then it’s a ripoff worse than whatever perceived transgression AI has created.
I can say this because I’ve had one of my images used by a world class professional athlete without being given the first credit for my work.
Edges cut
I thought about mentioning crediting image sources in my post but figured it was long winded enough. A good point to be sure.
Remember that the AI slop generator was trained on stock images without compensation to the photographers or rights holders. Though perhaps the relative scarcity of mountain biking photos in the training data is why Jeff's preferred slop generator couldn't quite work out what the fuck a tire is, so your point has merit.
This is the point that many forget when there's outrage over the use of AI for creative work. Those capabilities were built on the learning (mass consumption) of human-derived material with no compensation whatsoever. We haven't even touched on the fact that the output is now some gross generalization of individual creativity, so now our world gets painted with a single brush rather than many. This is why most low input A.I. generated images look the same depending on the model you use. Low input is an interesting point because we can also discuss the erosion of true skillsets (think excel formulas to build out a worksheet, dashboard, etc) and now we are left with how well can we give the machine directions. It's a shift in skillset building and problem solving, but to a lower quality tier. I realize that this likely sounds familiar to those that came before me and lived through the different eras of industrialization so hopefully I'm wrong and this is just a similar development of new tech packaged in a more modern vessel.
To bring this back to Jeff's post: why should we bother asking him for a breakdown on Fox's latest financial blunder when I can just prompt ChatGPT? It's because we recognize and trust that what Jeff has to say comes from a place of expertise, and that we appreciate his unique voice on the topic. When I see A.I. generated anything, I don't feel the same way.
You guys have done it. Jeff is surely going to hand paint the image for his next post. Thanks for making him see the light.
Post a reply to: The Bikeconomics (Mega)Thread