pigpop

We have all been tricked, unfortunately.

> people (arbitrarily, some could say) value the human production

This is a different argument, going into it would muddy the waters. It would devolve into a debate about how much and what type of input and individual perception. It's also been rehashed a million times since Duchamp thought to put a mass produced urinal on display as a fountain.

> most outputs being mass produced are of a substantially lower quality than people (and what is possible)

Different purposes and mass production does not eliminate the market for artisanal work, it replaces only the limited batch work done by hand previously.

> A problem is people are being tricked into it...

God help anyone who learns how the sausage is made.

> "if you can view this for free, it isn't meant to be sold. If it is sold, I deserve to have my time compensated"

This is mostly a matter for the courts but as long as the works are not being reproduced in whole unattributed or in violation of whatever license they held then it will mostly be allowed. If you look for it, you'll notice that artists "borrow" from each other constantly with no attribution and no payment to the originator. All but the grossest plagiarism goes entirely unpunished because everyone realizes that no one can stand alone. Is it different now that it is being made available as a service? Yes, there are now more legal ramifications due to an audit trail of scraping and purchases combined with the much larger output that can be measured and compared. Does it destroy the commons? No, people will use the outputs in their own work which will be published and become part of the commons just like any new style or media.

> that appears to be a "problem"

It's a solved problem. Synthetic training data is a valuable addition now. It requires a way to validate or rank it but that's not any worse than cleaning and filtering human data.

> conspiratorial thinking

An excellent smokescreen but we've been living in a soup of mass media propaganda for over a hundred years now so there's no denying that influence campaigns exist. Denying that humans act in coordinated ways to influence the opinion of others is absurd when you can get a degree in it, have it as a job title and build a trillion dollar industry on it.

How can the output both be so much worse than what a human could produce while simultaneously taking all the jobs and displacing the human made data it was trained on?

If everyone hates AI output and refuses to buy anything made with it then in what way is it a threat to people who don't use AI in their work?

Are these talking points perhaps crafted to benefit a small group of people who will lose the edge they currently have to a larger number of people enabled by these new tools?

And why do those millions of other people have that opinion? Have they arrived at it of their own accord?

Yes, you can create garbage with AI tools. You can create garbage using Blender, Unreal Engine, Unity, GIMP, Ableton, iMovie and Word as well and there are mountains of garbage created with those tools that nobody pays for or likes. If you have some talent and put the same effort into creating art with AI that you do with a digital paint program then it won't be garbage. AI isn't a magic wand that you can wave around and get perfect results, it can save you time just like using an image editor saves you a lot of time compared to drawing and painting by hand but it won't do all of the work for you. It's a tool and a poor craftsman that blames his tools for his failure.

My point exactly that big tech was loved for a while and then hated and then loved again and hated again. Have they truly changed that much that often or is it your perception that has changed?

Do you not remember when they were dragged in 2015-2016? or 4 years before that when they were heroes instead?

People are being led around by the nose. Cui bono? 2000 years and most people still can't manage to ask that question reflexively. AI hate is a parochial narrative being pushed by people who stand to lose influence. Examined objectively, there are reasons to be cautious when adopting AI tools but a knee-jerk rejection of everything AI is a purely irrational response that has been drilled into people through repetition, conformity and media influence.

Tech companies were never the "good guys" any more than they are the "bad guys" now. Those are just colored filters thrown in front of the lime light to shift your perception. As soon as tech or AI works to their benefit then suddenly they'll be the good guys again and no one will care they were the bad guys before. People just go where the algo leads them, believe what their influencers believe, think what their peers think.

> Admission rates at the most selective schools dropped from roughly 15–30% in 1980 to 3–6% today

They wouldn't be the most selective schools if they expanded access to everyone. They are performing their function and I would think the drop in admission rates would be due to the increase in number of applicants unless those numbers are already adjusted for that.

Doesn't the graph make this easy to interpret? College used to not be seen as necessary by the majority of people and so only those with a strong aptitude or desire to pursue higher education did so. It became a metric to use during hiring because a degree accurately indicated that the person holding it had high aptitude or a strong interest in the field and the competency to obtain a degree. People observed this and tried to game the system by acquiring any degree to pad their resume (the motivation changed from real aptitude and interest to simply getting a higher paying job). This worked for a time while there was still an overhang in the perception of a degree but eventually the sheer volume of people with degrees made it evident that it was no longer an accurate measure. The value of a bachelor's degree in any field plummeted and was replaced by other measures (a masters, multiple degrees, PhD, exclusive schools) which were also gamed by applicants. Now the jig is up and we're in a situation where the measure is whether your family has enough money and connections to get you a laundry list of qualifications or you rely on contacts or demonstrated talent to get the higher paying jobs. The widespread discounting of the importance of higher education is a good thing, it means that people are realizing that it is no longer an exploitable measure and will hopefully lead to a correction in both the number of people getting degrees and the cost of pursuing a degree. The unfortunate part is that we're in an interim period where costs are high and there are still a lot of employers that see a degree as a box that needs to be checked.

Yes, that's always the best approach whether you're directing an AI or a team of developers. One person has to have a coherent vision and the final say on what to keep and what to change. AI might get good enough where that input becomes minimal in a lot of cases was my point.

Hope I wasn't too harsh. It does look like you've accomplished a lot and it is the type of game I enjoy playing sometimes. A bit more work on performance and UX is probably all you need to have a finished game that would be fun to play.

That's an overly pessimistic take. An optimistic take that deals with some of the points brought up would be that after there is no need to earn money then addictive algorithms and dark patterns to drive user engagement will be successfully made illegal after being recognized as a type of harmful mind altering drug. People will still go to university but they will primarily study the humanities, especially philosophy. Some people will still study the sciences but more out of an interest in understanding some of what the machines are up to. People will still be involved in creating media, literature, performance, etc. right alongside the machine made or machine assisted media and art because people will always want something new. Even with 10 billion channels there will still be nothing on TV worth watching. There will still be the same distribution of motivation, sociability, curiosity and intelligence unless we start engineering those things genetically. People will do things just for fun or because it interests them, they do that now even though the chances of being world class at the thing are slim to none for most. Who cares if you're the 1,000,000th person to build an ultralight airplane and the manufactured ones are 1000x better than what you can build? It's still an interesting experience for you and a lot of fun to fly once it's done. There will probably be many more things to do than there are today because the market for them will be so much bigger. People will think of all kinds of unproductive things to spend time on and talk about. Sharing those experiences with other people will remain as relevant as ever whether it's 5 other people you know or an audience of 100,000.

Maybe this could be good but it needs a lot of optimization. A top down 2D game really shouldn't be that slow even in a browser. It's also very unclear what to do at the start. I only played for a few minutes because I didn't really know where to begin. My only comparison would be to Harvest Moon or Stardew valley where you're given some tasks immediately to get you started. Even a very short intro would help a lot to just get you started on the first game mechanics. Without anything to do a lot of people are just going to bounce off of it especially since moving around the world is so sluggish that it isn't even fun to just walk around and explore the farm.

> To be online, many now admit, is to be endlessly manipulated and exploited. In December, Oxford University Press declared rage bait to be the word of the year, noting that use of the term had tripled over the previous 12 months. The lexical spike represented a heightened awareness that monetized social accounts churn out bad-faith takes and extremist content.

The irony.

Yes but the solution didn't, though it seemed somewhat obvious after I read it (as is usually the case). Crucially, it also identified what didn't feel right without me specifying it.

All models bias towards the taste of the evaluation model which was built on the preferences of a group of people who did manual ranking of output. It results in a pretty bland design by committee style (though sometimes it works really well like with the early Claude models that seemed to have a coherent and interesting personality).

Currently, if you want better results, you have to prime the model with what I like to think of as a Context Mood Board consisting of material that you select to express a gestalt. Turning reasoning off also helps since it forces it to rely on what's already in the context.

This is a lot of manual work to set up but you can save it as something that gets loaded when needed. For open weight models you can create a LoRA based on a dataset that represents what you consider to be good taste.

The point is that it is possible to do even though it's currently not available from the major providers. They're more focused on maintaining their lead in capability, taste takes a backseat. At some point it will become more important and then we'll probably see either a curated selection of styles to choose from or they'll just keep them hidden and automatically route you based on your request and your user profile.

If you want a preview of the difference this can make try out some image generation models in ComfyUI and compare the results from the same prompt using the original base models versus the fine tunes and LoRAs created by the community. It's incredible what a difference it can make in the output and what you're seeing is the design sense or artistic taste of the person that did the fine tuning and the artists whose material was selected as input.

Looking even further ahead, there's a huge amount of undiscovered talent out there in people who lack the technical ability to produce art but have a very good sense of taste. Somehow getting access to their preferences would advance things even more. There's still a lot of training to be done so we're hardly at the limit of what is possible for instilling taste or artistic sense in LLMs.