jp57

I don't think this is the death of creativity. It is just a new kind of non-creativity. He is comparing to a baseline of zero (or perhaps negative) creativity before AI.

Without AI, it's likely that neither of these videos would have been made, right? Five years ago, was it a normal thing for even one, much less two, groups of wedding guests to produce a video to show at a wedding, and if so, would the video even have been good? I mean actually good, not "it's the thought that counts, A-for-effort" good.

Isn't this like saying, "I decided to write a song for my friend's wedding, and I used AI to help me, and I played it and everyone liked it, and then someone else played the song that they had AI write for them, and it sounded just like mine, and everyone loved that too!"

Writing songs is something that's been around long enough, that if guests had made a habit of performing their original songs at weddings, even a graybeard like me would have experienced it. The baseline to compare to is no original songs were performed.

As it happens, in the weddings I've attended, I've heard exactly one such wedding song, and it was awful.

I think one of the saddest thing is that the kind of person who would recognize, "we can solve this seemingly complicated problem by just applying this formula", would often have trouble even getting recognized in many corporate environments.

I managed a guy like that. He was capable of very complex thinking, but he wasn't in love with complexity, he was in love with simplicity. His solutions tended to be of the form, "we can ignore all these things, and just focus on X, and it will provide all the value." He'd notice something and simplify it and the benefit to the company would be measured in multiples of his salary.

Every manager who'd ever directly managed him knew what a treasure he was, but it was often hard for us to convince others of the value of his solutions because they were so simple, and people were convinced that hard problems must have complex solutions. (or else they would have solved them, right?)

He eventually got bored. He retired and joined a seminary.

What does it mean for an OS to be ported to another OS? Do they mean "ported to devices that support Android"?

I think this brings out the cognitive dissonance around "safety" regarding cyber security:

a) In order to make us safe, the LLM should help us find (and fix) the vulnerabilities in our own code.

b) In order for us to be safe, the LLM should not find vulnerabilities in other people's code.

I don't think this is resolvable in a way where both (a) and (b) win.

Not specifically about drinking water or jackets, but I've often wondered why air-conditioning condensate is plumbed into the sewer instead of someplace useful. An A/C is a water-from-air device, and they run in much of the year in most homes in the southern US.

The typical design that I've seen plumbs the main condensate drain into the sewer, and usually has an overflow that dumps somewhere harmless if the main line clogs. (like out of the ceiling over the bathtub or out of an exterior wall in a visible location, so you can see it and fix it)

The few times I've seen the overflow, it's been quite a fair amount of water. Certainly enough to help with garden irrigation, if nothing else.

I think he could have gotten away with naming his daughter Photophone, but he would have to pronounce it as if she were a character from an ancient greek epic: phoTOphoNEE.

It wasn’t wrong, though, in my case.

Actually seems absurdly simple now, but sometime last year I was trying to figure out what I'd need to tow my daughter's car cross country with my truck: what are the trailer/dolly options, what do they cost, can my truck actually tow the combined weight, etc.

I started out prompting ChatGPT kinda how I would with Google, one small prompt at a time, asking about various details. But after one or two of those I just tried "I want to tow a car of make A with my truck model B, from point C to point D, what are my options?" And it wrote me a report with comparison tables and computed towing weights and other details for different options.

At that point, I was like "Oh. This is different. And it's just the beginning."

Just recently it occurred to me that the sage parenting advice, "Don't try to make a happy baby happier," applies to so many other things. Once I had this idea, it seemed like everywhere I looked were people trying to make happy babies happier. Improving tools that work fine, optimizing things where the available margin for improvement is small, etc.

That’s far more believable than 10,000 elevator attendants. I was an adult in 1990 and used travel agents. But I can’t remember ever encountering an elevator attendant.

I really think a majority of NYTimes and ABCnews consumers don't know the difference between a 2/3 chance (super close) of winning and 2/3 of the vote (a landslide).

Wait. There were 10000 elevator attendants in the USA in 1990?

Well, he is standing still, the camera is stationary. For just that last segment, it is easy to answer "how did he do it?" Write out his remarks, rehearse with a timer, then figure out at what point in the countdown to begin speaking.

The main thing is that he has say basically one sentence right in a single take, but he is a seasoned television announcer, so that in itself is not too surprising.

The much longer segment, including walking with a moving camera at exactly the right timing, would have been much harder to get in a single take. (Not to mention that that Saturn V lying on its side is probably not even in the same location.)

They mention Meta's layoffs, which probably have more impact on employee morale than the AI stuff.

My current theory of tech layoffs is that over the last decade or so, churn-inducing practices like stack-ranking have gone out of vogue. One can speculate as to why this happened. Perhaps generational made middle management unwilling to do the dirty work? Nevertheless it happened.

However, companies still want to, and some would argue need to, eliminate low performers, so now they periodically do a companywide reduction in force and frame it with whatever justification is handy, macroeconomic conditions, AI, whatever.

This hypothesis would explain phenomena like companies hiring aggressively during or after a layoff, and why the layoffs keep happening year after year.

But are they actually serious people? I had corporate astroturf accounts arguing with me on my otherwise-ignored blog as early as 2004. All this time later, I just assume that every serious corporation employs PR firms using sock-puppet accounts to shill in favor of whatever dark shit they're doing, acting like it all just really great and good for us.

Claude is a mediocre programmer that can do great things with great supervision, but it can't make mediocre human programmers into good ones, because they can't provide great supervision.

It will try and try and try, though.

IDK. I was just reporting what Boris said. In any case the litany of reports of slop inside Claude Code speaks for itself.

There’s all kinds of stuff there that’s not the Claude Code app.

I saw a talk by Boris where he said, basically that Claude codes itself now. They have it automatically writing features and reviewing PRs, apparently. I suspect that much of the code has never been seen by human eyes within Anthropic.

Ha. Yes. "Speedrunning enshittification" is the phrase that's been in my head.

The flat-rate plans were the top of the slippery slope to enshittification, really. If everyone were on metered billing there'd be no reason for all these opaque and sneaky attempts to limit usage. People would pay for what they get and get what they pay for.

I have my doubts about the value of a two-alternative forced choice task for this. I was pretty much answering randomly both of the time because I wouldn't ncessarily have called either green or blue.

Tokens have replaced LOC as the dumb productivity metric of choice.

Not everything sells out right away, though. I've bought concert tickets on the day of the show more than once. Somehow they still managed to have all the concessions staffed.

But regardless, the formula for decreasing the price could be adjusted. For example, it could be an exponential decay toward the reserve price, with the decay rate set so that most of the decline in price is early.

Or, for shows that are entirely general admission, like festivals, you could use the alternative form of dutch auction: when tickets go on sale, everyone bids what they're willing to pay for some number of tickets. Then the bidding closes (with ample time for planning), and the bids are cleared in descending order of price, and everyone pays the amount of the lowest clearing bid. This method would find a price closer to the true market price of a ticket and discourage speculators.

Our basic findings suggest that the auctions “worked”: price discovery substantially improved; artist revenues roughly doubled versus the ­ fixed-price counterfactual; and, perhaps most importantly, the auctions eliminated or at least substantially reduced potential resale profits for speculators.... And yet, over the decade that has passed since the time of the data, rather than coming into more widespread use, ­ primary-market auctions for event tickets instead disappeared.... We conclude by speculating as to why the auctions failed to take off. As discussed in the introduction....

They don't seem to mention the most obvious reason: the same companies profit from both the primary and secondary market. Why would TicketMaster want to reduce the number of resales when it collects fees on them?

The horizontal control of venues is only one issue. A perhaps bigger issue is the vertical integration (if that's the right term) of first-party ticket sales and resale in one company. Ticketmaster has no real incentive to try to prevent resellers from buying up all the tickets on first sale, because it gets to charge fees on all the resales through its platform. The more times a ticket is resold, the better.

I don't believe a court would ever mandate this, but I'd like to see tickets sold by dutch auction: All tickets start off for sale at some very high price, like $10000, and the price declines by some amount every day until it reaches a reserve price on the day of the concert. Buyers can purchase as many tickets as they want, but professional resellers would have to guess the price that would let them clear their inventory at a profit. Under a system like this the best seats would go earliest (at the highest prices) while the nosebleed seats might still be available on day of the show, or not depending on demand.

This reminds me of the Plan for the Improvement of English Spelling by Mark Twain.

https://faculty.georgetown.edu/jod/texts/twain.html

Except that the mere existence of the market with the question posed for people to consider, probably activates the availability heuristic[1], causing people to overestimate the likelihood.

[1] https://philopedia.org/topics/availability-heuristic/

AI tools right now remind me of the old days of single-user PC/Mac operating systems without protected memory or preemptive multitasking. You could read any file, write directly to video memory, load machine code into the heap and then jump to it, etc.

IDK. The neediest possession I own is my sailboat. It is never finished. And it’s incredibly low tech. The highest tech part of it, the gps chart plotter, is not internet connected and it more or less works like that guy’s old watch. Meanwhile, the physical boat is desperately needy. An infinite to-do list. Other sailors i talk to say this is pretty normal.

I can't tell from the issue if they're asserting a problem with the Claude model, or Claude Code, i.e. in how Claude Code specifically calls the model. I've been using Roo Code with Claude 4.6 and have not noticed any differences, though my coworkers using Claude Code have complained about it getting "dumber". Roo Code has its own settings controlling thinking token use.

(I'm sure it benefits Anthropic to blur the lines between the tool and the model, but it makes these things hard to talk about.)