- Quoted phrase in the opening sentence
- Opening quote: “AI”
This is a small world. The US has something like ten major airlines. Eight if you don't count subsidiaries. Someone at the first one made a decision to embrace an exploitative pricing model. And then, someone at the second airline looked at this and thought to themselves, "you know, we're not gonna organize a PR campaign to stigmatize it, we're gonna try the same".
- Quoted phrase in the opening sentence
- Opening quote: “booking a [flight|hotel|concert tickets]”
- already streamlined (spend 2min on Google flights, seatgeek, etc...); we aren't stuck with early web UX and latency
- I want to have 100% control of the outcome or would have to write a long list of preferences and rules (middle seats, layovers, etc...) to feel good about trusting automation
- require making a large payment. the agent is not getting free reign to use my cc
- give me some personal satisfaction and enjoyment - its delightful to look foward to a trip, concert, etc... and planning is part of the fun and anticipation imo
If this is indeed the case -- good riddance to that practice! Why should I be sorry that it goes?
At this point it feels like I have an adversarial relationship with many merchants I use online. I wish the simple “you have a product, I have money” relationship was possible. I don’t want to take part in your giant human behavior optimization problem. You have a product and I have money why does it have to be more complicated than that?
Realistically though I know that’s not happening. I will continue to access online merchants in the way I prefer and will do my best to avoid detection or blocking. At least so far Claude is happy to play along but I suspect I will need to seek out an obliterated agent in the not to distant future. (Once the corporate capture of the frontier cloud AI models is complete.)
Just to be clear, your argument is that it's "counterproductive" and a "local optima" for people to have complete and accurate price information available for purchasing their tickets? And that making this information harder for them to find is better for them? Can you point to any real-world examples of an asymmetry in pricing information ending up saving the side with less information money rather than the side with more information to use that in a way that disadvantages the other side?
What if the airlines instead focused on delivering great service at a competitive price?
"Nah, let's escalate this arms race instead."
I was told later by a docent that my reaction was not uncommon. >smile<
The remainder of the valid Perl programs are mostly valid due to a Perl
language feature that is not commonly present in most programming languages.
Namely, Perl has a feature called “unquoted strings”, in which a sequence
of alphanumeric characters by itself is parsed as though it were a quoted string.
As an example, Figure 5 is read by OCR as the text ME, which evaluates to the
string “ME” even though the ME isn’t quoted. This would result in a syntax
error in most other programming languages.
(To be fair to Perl, when perl is run with the -w flag to enable warnings, it does
helpfully inform the user that at some point in the future, the Perl developers
will most likely pick gggijgziifiiffif as a new reserved word:
Unquoted string "gggijgziifiiffif" may clash
with future reserved word at - line 1.)- Quoted phrase in the opening sentence
- Opening quote: “you can just buy stocks too”
1. Boyce, Brown & Moore (2010), “Money and Happiness: Rank of Income, Not Income, Affects Life Satisfaction”
Exactly why the general public thinks Y2K was a "hoax" or "overblown".
Context: in 2012, I literally quit Fog Creek, largely over the fact that we had no backup strategy and I was told that a backup strategy was not feasible. I was "convinced" (paid a lot) to stay on for another two years to see if Kiln would succeed. On my way out of that meeting, I commented to a coworker, "I mean, what's the chance a hurricane hits Manhattan in the next 24 months?"
100%. The chance was 100%. And the only reason people had to haul open containers of diesel up 17 flights of stairs was because no one wanted to admit that the chance wasn't 0%.
Anyway, not sure what was going on with Jane Street here, but holding up that Fog Creek/Trello/SquareSpace story as a positive always drives me somewhat nuts. It lionizes a failure.
It took me an embarrassingly long time to realize that (my 30's), and now I keep my conversations with my coworkers very utilitarian and divulge minimal information about myself. I try not to be a dick or anything, I'm polite enough, but pretty much everything I talk about is either directly relevant to work and/or so surface level as to border on "vapid".
ETA:
Just as an example, I used to have really bad sleep issues in my 20's that would make it so I wouldn't be able to get to sleep until around 4am, if at all. I made the mistake of bringing that up at work to my manager. Two weeks later I got a phone call at 3am because there was something he wanted to look at and "he knew I would be up". I was up, but I didn't really have much desire to get my laptop out and do work. I liked that job at the time and I liked my manager, so I did it anyway, and it wasn't until years later did I think about that incident and how utterly weird it was that I was sort of being punished for my insomnia.
My sleep issues have gotten a lot better now, so if someone calls me at 3am I probably won't hear it because I'll be asleep, but I remember that incident now and have it constantly spinning in my brain when I chat with coworkers.
- Quoted phrase in the opening sentence
- Opening quote: “low voltage”
Bo, this looks very cool, but I'd encourage you to immediately add "Based on Papers Please by Lucas Pope", with a link here:
- Quoted phrase in the opening sentence
- Opening quote: “oh yeah, this is a useful thing that we should have in our product portfolio”
But maybe this is one of those things that's only obvious in retrospect? I'm sure people were trying something similar years ago with grammar-constrained outputs (for example to only output a simple json structure), but you can't get probabilities out of the model that way. Which makes me wonder: how do these models calculate probabilities?
From my own experience, these learned methods from Apple, Spotify and others, are highly relevant to hyperscale only - when you are dealing with hundreds of millions of items to choose from (or more), and require super high relevance in the top-5. For all other use cases (up to a few million items) I found that just building a BM25 index, and giving the LLM a BM25 "search tool" and casting a "wide net" wins every single time compared to learned methods (though the latter are more publishable in recsys and icml).
[1] https://research.atspotify.com/2025/9/you-say-search-i-say-r...
- Quoted phrase in the opening sentence
- Opening quote: “We made a determination that we have met this threshold,”
We have made a determination that more hype is needed
My goal was not to imply that we're not shipping faster than we were prior to AI, as I think that's beyond debate at this point. What we investigated is why we're not shipping faster than we are today. If (high quality) code is being written at five times the rate it was before and we're shipping "only" twice as fast, then what's the reason for the difference? The talk shares results from studies that help improve our understanding of why.
However, for those that are actually curious, the submission is not itself a cynical or dismissive take. The question in the title is interrogative and contextual, not rhetorical and universal, and the presentation tries to break down some of the bottlenecks and challenges that are leading to AI integration having a different impact on software engineering output than some people might expect.
Rather than suggesting those bottlenecks/challenges are universal or intractable, it just enumerates some of them and presents some survey analyses on how widely encountered they are.
It's actually pretty interesting and the kind of thing that engineers and those who work with them want to be reflecting on, so that they can better engineer their usage and processes around AI itself.
- Quoted phrase in the opening sentence
- Opening quote: “The Mythical Man Month”
"One occasionally reads newspaper accounts of how two programmers in a remodeled garage have built an important program that surpasses the best efforts of large teams. And every programmer is prepared to believe such tales, for he knows that he could build any program much faster than the 1000 statements/year reported for industrial teams. Why then have not all industrial programming teams been replaced by dedicated garage duos? One must look at what is being produced."
Brooks argues that the coding that happened in the garage is only a fraction of what you need for the final product:
"It costs nine times as much. But it is the truly useful object, the intended product of most system programming efforts."
What we're seeing right now does not feel much different. You can greatly accelerate a first draft, a greenfield project, an experiment, or at least the initial iteration of one. Changing a live production system or a widely used library is an entirely different proposition.
- Quoted phrase in the opening sentence
- Opening quote: “closed”
There are reasons why folks know the names of notes in some positions better than others.
When I was playing a lot of bass for folks in low rent with bands that were playing standard music and never rehearsing, I generally only paid attention to the low string and worked my parts out (often on the fly) around the relative pitches.
If you're trying to keep up with the bass part on a song you've never heard before while performing it on stage, it's much more relevant (and far easier/faster) to know the position of the minor 3 note of the scale than it is to know what note is the minor 3.
Over the decades I play a lot more piano, and that changes. I play a lot of blues and jazz but am quite poor at reading typical rep for the instrument. While I started out mostly playing using chord fingerings (and inversions), the more I play the more specific I can get on my understanding of actual notes. Because I've been playing a lot of clarinet (which forces me to transpose a whole lot, including reading a part and transposing on the fly), the names of specific notes has gotten a whole lot more important, but also easier with practice.
On the pedal steel (which I have only been playing for a decade on and am still coming to terms with), while I can pull all that knowledge I am still usually playing based on a notes position- I know what string/lever combo is going to get the 6 note in a given scale rather than thinking "I need to play an A to get this C chord to be a C6, and that's this position on the neck with the BC pedals down or that position with the LKR lever pushed.
All that is to say, it's okay to not know the names of notes if you understand their function related to the melody.
Maybe he just said that to make a sale, but yeah the fact they offer that targeting alone is quite telling.
- Quoted phrase in the opening sentence
- Opening quote: “LG TVs do not…”
LG: No. ACR in LG TVs does not collect or store screenshots, screen images, screen recordings...
Hold on, we already know this is not true. There is even an option in the settings to disable this (which is usually automatically reverted by LG within a week).The setting has the obvious name of "Live Plus".
- Quoted phrase in the opening sentence
- Opening quote: “no answer”
It is much more of "let's sell cards to crypto people, but nerf them, to not harm our enterprise margins", not "we've got a ton of subpar hardware, let's to something with it".
I do, however, see a massive difference between "AI as tools" and "AI as creative director". I agree with the mindset of "I want to hear stories from humans" and "I want to experience creative expression from humans". But some of the current discussion feels like the arguments we had back in the day of whether Photoshop/digital media/video games were "real art" or not. If a human has creative vision and uses a robot stir-er to mix their paints, I don't think that should invalidate their painting. We can still judge it, though.
"Prior art has reduced the novelty of the discovery" is a great one
The Walt Disneys of the world screwed themselves on this one. If copyright was still the 20+20 it started as, and if the bulk of people being "stolen" from (don't get me started on using theft to refer to IP), then I think most people would be much more up in arms over the AI industry.
But, in a world with sane copyright laws, perhaps AI companies would have different approaches. Today, if they want anything after 1930, they're either breaking laws or in uncharted territory.
And siloing our knowledge is a silly response even if you are a copyright maximalist.
Perhaps google SWE should have requested a literature search before publishing his results as novel
But Dr. Estrada-Belli found something that was undiscovered. Finding this structure wasn't his number one goal - it was the side effect of a larger anthropological assessment that allowed him to improve estimates about population density of the north versus south, how Mayan homes and farms are distributed, what this means about social relationships between the elite and common people, how different regions were interconnected, etc.
Interestingly, one of Dr. Estrada-Belli's primary conclusions is not that these structures were only used by "the local elite," but that the closeness of residences and farms to these community sites suggests that the elite and the common people actually had a much closer and nuanced relationship than we previously thought. That's the opposite conclusion you'd get from the Google SWE's assessment.
Knowing that a Mayan site exists is one thing. Understanding what it means is another.
- Quoted phrase in the opening sentence
- Opening quote: “This is by far the best time ever to be alive.”
I appreciate the enthusiasm. I'm not sure I agree. Especially since this was a SWE and not an archeologist / anthropologist announcing this (especially if the goal is to understand and preserve history -- not get headlines for AI).
The final barrier for super intelligence, or perhaps even just "human intelligence": can it materialize "new fields/ideas" without large datasets of human made and curated existing attempts?
Absolutely not. Constraining the set of possible actions instead of having you verbalise what you want every time is often a feature, not just a technical limitation. A camera with a shutter button is better design than having to tell the camera “please take a photo”.
At the supermarket, I personally also prefer the self-checkout machines over going to a human cashier.
Commodities like todo apps and calculators rarely compete on features, they compete on product design. This includes look and feel, relation between elements, design choices on what exact common features are available.
All products are, at their essence, just a set of design decisions stitched and shipped together.
Now, making these design decisions is hard. It requires knowledge, consistency, and insights into how real people operate real products. Even as a sole user of the app you build you rarely know what you actually need.
LLMs are 'design-less' in their essence. Building your own todo app takes literal seconds and it all looks simple because LLMs will default most of these design decisions for you - until you try to use your new app daily and notice it somehow sucks. Building a good todo app or good calculator requires tweaking, publishing, polishing, and solving bugs, testing, rinse and repeat.
Another tricky side of this is that you might know logically what features you want, but you are not selecting one product over another logically. You need to try and see if that works for you, all the vibes and good parts and bad parts and design choices and colors, till you find something that "clicks" for you. Using pure Claude as a todo app only works as long as you want to have more Claude in your daily life.
In the meantime app store provides dozens of well-reviewed alternative solutions which look good and require zero maintenance.
And that's just a calculator or todo app. Non-commodity apps are even harder than that.
- Quoted phrase in the opening sentence
- Opening quote: “Harbor is currently in a private preview under Apple’s Claris brand”
There's a name I haven't heard in a long time! Claris (always an Apple subsidiary, since being founded in 1987) renamed themselves to FileMaker in 1998, but apparently they rebranded back to Claris in 2019.
Looks like they've mostly just released annual upgrades to FileMaker for the past decade, but they did add MCP to it last year: https://www.claris.com/blog/2025/why-mcp-matters-for-claris-...
The Apple and Microsoft community support forums may have had more value years ago when the only people who knew how to find and use a forum were already technical. It’s hard to imagine now, but there was a time when making an account on a site to talk to other people was unusual for average people.
Today, the only useful community support forums I visit are for highly technical products. Anyone participating has a baseline level of competency to even get into the software or hardware, so conversation quality is higher.
The Apple (and Microsoft) community forums almost always have a 12–page thread of people having my same issue and no solution. There is also the strange phenomenon of the community members who want to be very helpful but are only copying and pasting “have you tried turning it off and back on again” or “wipe your device and reinstall” non-solutions.
I’m surprised the Apple community forums lasted this long. I can remember a couple solutions I found there, but most of my visits were frustrating experiences of clicking through pages of threads of people complaining about the same thing. Even when someone did post a solution, it could get lost in infinite posts from users who didn’t know how to use a forum but kept adding complaints to the thread.
- Quoted phrase in the opening sentence
- Opening quote: “hey this thing is broken”
- Opening sentence uses a quote to challenge or distance a term
- Opening sentence challenges the term “hey this thing is broken”
The Apple and Microsoft community support forums may have had more value years ago when the only people who knew how to find and use a forum were already technical. It’s hard to imagine now, but there was a time when making an account on a site to talk to other people was unusual for average people.
Today, the only useful community support forums I visit are for highly technical products. Anyone participating has a baseline level of competency to even get into the software or hardware, so conversation quality is higher.
The Apple (and Microsoft) community forums almost always have a 12–page thread of people having my same issue and no solution. There is also the strange phenomenon of the community members who want to be very helpful but are only copying and pasting “have you tried turning it off and back on again” or “wipe your device and reinstall” non-solutions.
I’m surprised the Apple community forums lasted this long. I can remember a couple solutions I found there, but most of my visits were frustrating experiences of clicking through pages of threads of people complaining about the same thing. Even when someone did post a solution, it could get lost in infinite posts from users who didn’t know how to use a forum but kept adding complaints to the thread.
- Quoted phrase in the opening sentence
- Opening quote: “hey this thing is broken”
- Opening sentence uses a quote to challenge or distance a term
- Opening sentence challenges the term “hey this thing is broken”
> That would be a beautiful study, but it wasn’t the study that was actually conducted. The “lottery tickets” in question were in fact football cards, each one with a picture of a player, their team, etc. Some participants were given the opportunity to choose their football card and some had a card handed to them. Later, just before the drawing, all of the participants were asked how much it would take to buy their card back. On average, the pickers asked for $8.67, while the not-pickers said $1.86.4
> I’m sure you can see the obvious and less exciting explanation for these results: when people were given the chance, they picked their favorite players. The people who didn’t get to pick ended up with a card they liked less. This “illusion of control” effect is merely a “reality of people having preferences and acting on them” effect.5
Alternatively, people who are knowledgeable of a subject will have selfishly biased opinions about it, which they value more than unknowledgeable people will value an arbitrary stance on the subject. From experience, this seems to affect things beyond situations of chance like lotteries, and infests collaborations like businesses in ways such as Not Invented Here Syndrome or generally poor management choices that fail to consider situational factors (e.g. overarching decrees that are ineffective in certain situations/locations/applications).
The interpretation here is not so much that an LP can do Astra's work -- after all the Astra prompt was "improve Qwen's math by changing weights" or similar. Instead, it's that with some guidance (eliminate this behavior, preserve that behavior, as expressed through linear constraints) the solver is much more efficient than Astra's trial and error. So essentially the solver would be a way of cutting through all the time consuming stuff and expressing Astra's intent more immediately. And then presumably Astra could iterate on that to a better solution.
The use of an LP solver here is a really cool insight.
- Quoted phrase in the opening sentence
- Opening quote: “author”
I expect the problem isn’t necessarily the AI generation itself, but the difficulty to distinguish between someone that’s really understood the problem domain versus someone with LLM-induced “Dutch courage”.
That in combination with a large uptick in volume of community interactions (issues, PRs, …) risks burning out OSS maintainers.
Perhaps giving it a few months/years for the flurry to die down, and for people to establish better best-practices for using these tools is the motive.
Elixir is well handled by LLMs at this point, and we keep the “easy to scale and maintain” aspect (operational characteristics of the runtime are very good).
I can safely say that total cost of ownership is excellent.
My first pass at it was a naïve search using the built in string matching. Because of the size of the file, and the speed of PCs, it would sometimes take more than 30 seconds to do a search. Ward wanted it faster, much faster.
I eventually came up with an algorithm that took the text you wanted to search for, and changed it into a special kind of bit mask, table. It was a table 256 bytes long, initialized to zeros. If the first character of the string you wanted to search for was a D, then the first bit of the $44 and $64 mask would be set, to handle case sensitivity. A wildcard * was only 1 character long, but would set all bits for that character. I then used the 8088 XLAT instruction and a very small loop ingest a character of the text being searched, use that as an index into the mask table, and AND that with the existing flag in a different register, then shifted it, checking for CARRY (a match!) and looping if not zero.
It took search down to less than 3 seconds if the whole file was loaded in RAM. The thing we used as a benchmark was searching for "DOS TECH REF", the IBM DOS Technical Reference Manual.
I can't find the code, lost it over the years, but it was one of my coolest hacks.
A lot of the Security Framework came from Classic, hell, the idea of a keychain came from Classic! Just a mess.
When I installed WireGuard on macOS, I was happy to discover that it stores private keys in the system keychain. Accessing the private key requires Touch ID.
Then one day after a system upgrade, the WireGuard app crashed on startup with:
[APP] Unable to open config from keychain: -25300
Then the app deleted all profiles because they are all "orphaned", apparently for "keychain consistency" [0].From then on, I just use a text editor to edit my WireGuard configuration files, and import them to the app. The bug still exists to this day [1].
[0] https://github.com/WireGuard/wireguard-apple/commit/dfb685f2...
[1] https://www.reddit.com/r/WireGuard/comments/wyt828/osx_wireg...
I recently had a gig collecting soil samples in rural Alberta, and we relied on Starlink to upload the data as we worked.[0]
It was surreal trundling through a remote farmer’s field in a skid-steer with a funny-looking square dish on top, surrounded by a herd of bison or cattle.
Stepping outside of the machine to take a piss on a tree and hearing the motors in the dish whir as it repositions itself to get a better connection was startling and neat. It’s one thing to read about the dishes having that functionality but another to experience it while you’re doing something as mundane as a basic bodily function.
Pulling the dish and the mount off the skid-steer and tossing it in the back of the truck, and seeing the off-the-shelf Ryobi battery and inverter that powered it, all scuffed up and covered in dirt, makes you realize that as magical as it is right now, it’s just a machine. People thought that transistor radios were pretty magical when they came out but eventually they just became an everyday thing.
Soon enough these things won’t be so big and expensive, and there will be a market of providers offering even better service. What comes then will be scary from a privacy perspective, but it may also enable so many positive things that we can’t yet imagine.
“The most profound technologies are those that disappear.”
I just checked a few weeks ago and “Roam” does no longer include international coverage at all. Only “Roam unlimited” does, and only for 30 days at a time which entirely defeats the purpose for me. Also no way to replenish this quota or buy additional, you have to go back to your country for it to reset.
My SIM card already gives me 14 days of roaming included, so Starlink is just marginally better for my use case.
I’m sure it’s regulations related but still a bummer :(
- Quoted phrase in the opening sentence
- Opening quote: “absolutely cannot let Elon maintain a monopoly”
Those are the three options for your mandate.
the non-government intervention option is just for Starlink to prove the market, and others can move in. SpaceX is willing to sell them the ride for a decent price, in relative terms.
Agreed. My fairly obscure company got hit by something like this last weekend. What stood out was the speed: attackers found something small, used it to uncover something else, and pivoted to clever exploits for the new thing within minutes. It hit a dozen of our apps in parallel within the span of a day, from a rotating pool of residential proxies, Tor, GCP and Cloudflare IPs.
I'm fairly confident it was LLM-driven. The payloads and throwaway accounts had names like "PT59553 TEST IGNORE" and "PocCorp1659r2", and it adapted to error messages the way a person reading them would, just never sleeping. A non-tech company with 100x the surface area and much more valuable data could be in big trouble.
Of course I used LLMs for the defense too, and they were amazing at log forensics and fixes. The annoying part was when I asked Opus 5.5 for clear repro cases of exactly what the attackers did, I hit the cyber safeguards. Opus was trying to keep the exploits for my own bespoke applications away from me. It was happy to fix things of course, but I worry if it really can't 'think' about actual exploits, its fixes will be less informed.
> almost none are being exploited
Isn't the whole point of glasswing to patch stuff way before it can be exploited? If the patched code is distributed quickly- there's no opportunity (or incentive) for bad actors to craft and distribute an exploit. This reads like an attempt to spin them as not being "real" vulnerabilities.
- Quoted phrase in the opening sentence
- Opening quote: “super”
That's consistent with my experience using them for coding, where even the best models mostly work by being persistently midwit (with infinite knowledge) more than individually brilliant.
One corollary of this is probably that any project that has had an army of actual monkeys pounding on it for years is likely more resistant to LLM attack than one that has not. Everyone is worrying about individual robot geniuses finding obscure vulnerabilities in Linux, PHP, Python, etc., but they should probably be more worried about the permutations of less-known software floating around their systems.
The worst part is being thorough takes so much more work - they often do raise valid problems but you just can't set them in a loop because of this tendancy so you end up with way more work and hard questions
The history of general-purpose massively parallel machines isn't good. From the Illiac IV to the Playstation 3, there's been dud after dud in this space. You can build them, but the users don't come. This sort of architecture is great if you have some specific problem that parallelizes, such as finite-element analysis or neural networks.
(I've used an NCube, and I kind of liked the Transputer concept. But both were dead ends.)
Pitchbook shows it as "Out of Business" since 2019.
Of course, the “don’t update after this firmware version” bit would always be the case (since any widely-distributed such vuln will be patched eventually), but deliberately moving the timeline to the left is sad.