alex0015

The first time I watched it, I was a freshman in college and I could barely get through it. I hated the sense of awkwardness and what read to me at the time (2012) as intentionally cringe humor.

The second time, I had graduated college, seen more of the world and how people treated each other, and I felt like I got it a lot more. I didn't feel offended by the long awkward moments anymore. It was entertaining and I was able to see the virtues and flaws of the characters, and even recognize traits of myself and other people I knew.

I can't remember for sure whether I've seen it a third time, but I think I have, and it would have been in my mid or late twenties. With that watching, I was able to appreciate it most as a film and a product of its time, and imagine how it might have landed to its audience over the years. My recollection of it now is of someone who is by nature a pretty passive person struggling with the concept of his own agency, which speaks to me personally.

If someone took a close-up photo of a statue's stone texture, I don't see any problem with someone editing the photo and selling it as part of a texture pack for 3d modeling. I feel like what was done here involved a roughly similar amount of transformation. I agree that the right thing to do is attribute the source, though.

I remember this well into the 2000s. My parents have never stopped subscribing to their local paper and your comment made me realize how much advertising is gone from it actually. I remember enjoying reading the Best Buy, Target and Circuit City ads each week, especially the Sunday circulars. That was only a fraction of all the paid ads in the paper back then, too.

I'm still a codex man myself but lots of people feel like opus 5.5 just gets it for some projects in a way that astra doesn't. Especially given the fact that astra and sol 6.1 are often at capacity late at night during US hours, when lots of people want to run longer unsupervised projects while they sleep. There's a general sense I've felt reaching some comment threads that openai failed to react with something better than opus 5.5 and are rushing to release small tweaks to make people feel like they're still getting something new.

Interesting framing. I thought that was part of the appeal. When I was in college I was actively looking for an opportunity to go to another country and do a temporary work thing where you stayed in a mountain resort and worked at the front desk all summer.

The person you're replying to was presumably talking about a vacation town in the US. Double taxation, by the way, is only an issue if an American makes over $100,000 per year in their foreign job.

Great counterexamples! These are specifically part of the game, whereas in Chinese culture it seems to be more like "let's play a defined game with winner and loser, and in order to motivate you not to be the loser, we'll agree that the loser do something embarrassing." Like in the US when I would play recess games as a kid, if you lost, you just lost with nothing at stake. Also we're all in our thirties instead of kids playing truth or dare.

I (an American with a lot of interest in China) was playing doudizhu with some Chinese friends recently. It was agreed that the losers would have to kowtow to the winner instead of paying money. I was on the losing side and it meant nothing to me, while even in a private, joking setting it was apparently a very uncomfortable thing to have to do for my teammate. See also the general cultural attitude toward doing something embarrassing as a "punishment" for losing a bet or game, something fairly unknown in the US.

I worked in the university archives at Cal Poly more than a decade ago. They mostly had me digitizing and cleaning up scans and similar types of things, which was the best job imaginable at the time. I used a reel-to-reel tape recorder connected to an iMac and I'd just listen to hours and hours of 1970s campus radio in real time as it ran. It was all very high quality (in fidelity, if not production value) and the slow speed of the task gave me a lot of time to think about what life was like at the time the broadcasts had been done.

I have a clipper card and take VTA regularly but not enough to make a pass worth it. Commuting 3x a week, I may as well be paying with a Visa.

A friend of mine in high school (2012) was required to add to the notebook his parents kept in the car whenever he got gas. I wonder if his parents still do that kind of record keeping now that the car tells you its MPG, gas price historical data is easily accessible, and card purchases are tagged as "Gas" automatically in the banks that I use.

I do actually audit each transaction in my card statements each month to look briefly at the categories of spending I use most. Gas is a category, but that number doesn't really change my behavior. It's just a data point.

What ads do you see for channel membership? I run a channel and put some of my videos behind a membership gate. I mention the premium videos in some of my free videos but otherwise don't see options in the YouTube UI to actively promote my membership in any other way. Is it putting graphics or something inside the video to tell you to click the membership button?

I'm a mac user but not really a power user. Is there not some hidden system directory you can go delete yourself and be done with it?

Me too. I was just thinking recently about updating from Sequoia, now I've got to take a while and reconsider everything.

Yes but I'd actually go even further than that. We've all had experiences where the model straight-up produces gibberish sometimes, right? It's happened to me with a badly configured harness on a local model, and even also on frontier models like when you used to ask them something innocuous about the seahorse emoji.

What I'm arguing is that yes, it's your fault if you misprompt the model and it does something catastrophic. It's also your fault if you prompt it correctly and it does something catastrophic anyway due to some other glitch beyond your control.

The whole discussion started with the point that OpenAI could be financially responsible for damages if their models cause problems for users. If my job is to design a system that works, and instead my system doesn't work, it's my fault regardless of whether I used no LLM, a local LLM, or OpenAI's LLM. Depending on whoever's in charge of doling out consequences, I might get away from it with zero, light, or heavy consequences. But at all levels, I can't reasonably expect to deflect blame onto the model itself.