I have a contrarian take on this: if most people working on the codebase are using the same model, then these comments serve a very useful purpose: they tell the model's future self that despite its tendency towards this solution, it is incorrect and here's why. It encodes useful context right when the model needs it next to avoid making the same error.
I imagine this style of commenting is partially responsible for model improvement in SWE work, a trained trait through feedback, since it makes the comments actually useful. Not to you or me, but to an LLM.
Some of these errors are obvious to a human, that's why we find the comments unnecessary. But if everyone's using LLMs and said model tends to make this error, then they are very useful. It's like a comment you would leave for another human to help them avoid a footgun. It's just a LLM footgun, not a human one.
I agree, but that information should be in the commit, which the model also has access to. Of course it's more token intensive to look through commits, but "good" models might know when some more research about a particular bit of code they find weird would be a good idea. Thoughts?
The problem is that the list of things one shouldn't do is infinite. The things one should do are narrow and get narrower along two independent axes of clarity/wisdom: system requirements and engineer seniority. LLMs overindex on any words given, so you only want to give them negative guidance around truly repeated, almost common-sense pitfalls. But in a large distributed system often the changes come from all different angles, and each agent will find it's own unique failure modes. Those comments will dilute critical context for diverse agents far more than they will help on average, at least for the systems I'm working on, given current claude code chattiness.
One way to clear some of the low-hanging slop is to just have a separate agent code reviewing and pruning any comments that don't stand on their own purely in the context of the diff, but it still doesn't catch it all.
If LLMs are any good, like many of you claim, they won’t need these comments. They will ingest the code and know what was done. They won’t need comments about what wasn’t done.
Shouldn't a senior developer strive to eliminate complexity while increasing velocity? The two do not contradict. Reducing complexity can increase velocity.
It's probably there not as a way to connect networks, but as a way to keep them separate, only allowing RDP between specific computers on different networks.
A 15 frame min anf max GOP size would do the trick, then you'd get two 15 frame GOPs. Each GOP can be concatenated with another GOP with the same properties (resolution, format, etc) as if they were independent streams. So there is actually a way to do this. This is how video splitting and joining without re encoding works, at GOP boundary.
In my case, bandwidth really mattered, so I wanted all one GOP.
Ended up making a bunch of patches o libx264 to do it, but the compute cost of all the encoding on CPU is crazy high. On the decode side (which runs on consumer devices), we just make the user decode the prefix many times.
You're right, but the way to achieve this is to allow the error to propagate at the file level, then catch it one function above and continue to the next one.
However, LLM generated code will often, at least in my experience, avoid raising any errors at all, in any case. This is undesirable, because some errors should result in a complete failure - for example, errors which are not transient or environment related but a bug. And in any case, a LLM will prefer turning these single file errors into warnings, though the way I see it, they are errors. They just don't need to abort the process, but errors nonetheless.
> And in any case, a LLM will prefer turning these single file errors into warnings, though the way I see it, they are errors.
Well, in general they are something that the caller should have opportunity to deal with.
In some cases, aborting back to the caller at the first problem is the best course of action. In some other cases, going forward and taking note of the problems is best.
In some systems, you might event want to tell the caller about failures (and successes) as they occur, instead of waiting until the end.
It's all very similar to the different options people have available when their boss sends them on an errand and something goes wrong. A good underling uses their best judgement to pick the right way to cope with problems; but computer programs don't have that, so we need to be explicit.
I believe it's because if you are easily replaceable, then screwing up means you're not worth the trouble. If you aren't easily replaceable (whether it's because you have demonstrated you're a good employee or you're working a high-demand role), you are worth the trouble and you'll get more chances. There are other reasons too, such as jurisdictions where suing after being laid off is more common, which makes more chances, PIP and severance packages more likely.
>I believe it's because if you are easily replaceable, then screwing up means you're not worth the trouble.
These ideas work great when you have a large/growing labor population, we're seeing it start to fall apart in a tight/shrinking labor population. "nobody wants to work" is the drum beat of the employers that used to burn employes.
Well, then they're not easily replaceable. Which implies those workers have other options. And thus employers that treat workers worse than other employers will see workers leave and/or find it harder or more expensive to hire. So it all works out.
Well, maybe fast-food restaurants were a low-interest-rate phenomenon or whatever the equivalent term is (a high-labour-supply phenomenon?). If that kind of business can't afford good enough working conditions to get a decent supply of labour while selling its products at a decent price, then yeah, the whole industry will burn. But if there's a demand for it then sooner or later someone will start offering it. I think that probably looks like a decoupling between middle-class-oriented (e.g. Costco or Chipotle) and working-class-oriented.
During COVID a new kind of food business sprung up called a virtual restaurant (or a ghost kitchen), where instead of having to rent a restaurant space and hire wait staff, they just had a kitchen and delivery drivers. Lower expenses, fewer staff.
Eventually, those delivery drivers will be replaced by delivery drones.
I promise I’m not that whiggish when it comes to automation, but there was a time when a good portion of human labour was washing clothing, and now that’s become much less of a thing for much of the world.
Perhaps food service and in-person retail will start to go that way too. It’s my hope we can navigate that and still make a place where it won’t be so bad.
Sometimes we do our own laundry, sometimes we pump our own gas, and sometimes we prepare our own meals. Not having a servant to do those things for us doesn’t make us destitute; it makes us human.
Why don't I cut my own hair? Because I'm shit at it.
If you went to a barber, and you sat down for a haircut, and they handed you a pair of clippers, would you go back? Fuck no!
The problem with having customers do their own checkout isn't that the fat and lazy customers have to get off their fat asses. No. It's that customers are unbelievably shit at that.
Have you noticed that, despite there being, like 10x as many self checkout lanes than before, the lines are longer than ever? Go ahead and do the math on that. Not only are people doing more work, they're paying more for it too, and it's a much worse service. Lose, lose, lose. Unless you're the corporate overlords.
You may find yourself in that situation if you have a device that only supports SIMs, and you can't use any of the cheap travel esim providers with it. For travel, you would replace your local SIM with the 9eSIM, and be able to switch providers depending on destination. The difference can be huge in some countries, where a local provider's travel plan can be 30 to 50 USD, while a equivalent on an ESIM provider is just $4.
I live in such a country and have parents with older phones who can't use esims, so the value is obvious to me. :)
In which countries are eSIMs cheaper? I have never encountered this in Africa or in Asia. I was just in Vietnam, a local SIM was probably 50% cheaper than anything I could find on esimdb.
Currently I'm in Georgia, unlimited internet for a week is 9 GEL, or around 3ish USD per week. The cheapest on esdimdb is 19 USD for unlimited internet for a week.
What we usually do when we travel is buy the cheapest eSIM, usually on some introductory offer to get like 1GB for 1 USD (so we can order taxis, maps etc), then go to a local provider and get a local SIM.
One place where an eSIM was a good choice was China. I don't quite understand how it works, but it seems if you use an eSIM in China you get around the great firewall without needing a VPN.
I wish eSIMs were cheaper, so I wouldn't have to deal with the headache of doing that. When going to local providers, sometimes they offer an eSIM option, but there is usually no price difference.
> I don't quite understand how it works, but it seems if you use an eSIM in China you get around the great firewall without needing a VPN.
That's just the default for most mobile data services, eSIM or physical SIM. Your home network provides Internet connectivity. "Local breakout" (where you get an IP of the visited network) has never really taken off for various reasons, one being that people actually like being able to access everything they also can at home.
I also strongly suspect that this is why iPhones in China don't have any eSIM capability.
I don't know if eSIMs are more expensive or less than a local SIM, but they are much more convenient for me when I travel. I can have working internet as soon as I step off the plane, which is great for finding transportation and not having to deal with kiosks that won't speak any language I know and might be closed. I don't have to hand over my passport to get a SIM, and in China they get around the firewall. The cost of an eSIM vs the cost of travel is too small to notice but the convenience is always noticed.
Currently traveling, and the savings are real. Although in this case it’s the opposite: travel eSIMs rates are about $80 for data for 30 days, whereas a cheap local prepaid SIM card is $8-16 (but with no eSIM option)
The service is called Holafly, it was advertised on the plane, and my travel mate bought it without hesitating (because of the convenience of an eSIM, even though they didn’t get a local number)
That's basically guaranteed to be overpriced. Anything prominently advertised means you're going to be paying for the advertising budget.
Also, it's "unlimited data", which probably makes it more expensive than it needs to be due to adverse selection. For instance it charges $50 for 15 days in europe, but on esimdb[1] you can easily find esims for just over $1/gb. It might still be worth it if you're using absurd amounts of data, but citing it as an example of esims being very expensive doesn't really make sense.
I don't know about europe but 40 to 80usd for 15GB for 30 days in Mexico is completely crazy when you can get a physical telcel sim card with 25GB and unlimited data for whatsapp and all major social medias, which means you can easily go for the smaller 10GB (15usd) or 7GB (10usd) choice if the most you will transfer is on social medias and whatsapp.
I found my google maps app for navigation and image translation via google lens/translator app ate up a ton of my data. I had to turn off off a setting or two to reduce the maps data.
In my experience, getting an eSIM is usually cheaper than the airport SIM card plans, but often there are cheaper plans available when you get to the city. In any case, having both options is nice!
I think it depends on how you define “developing countries”!
I’m currently in Thailand and I got 6 GB for about $2 total (50 baht for the SIM with 3 GB for 3 days, another 20 baht top up for another 3 GB / 3 days). I did use eSIM for about a month before that though (I just wanted a local number to order some stuff from Lazada).
Another example (also from Southeast Asia, FWIW): Malaysian SIMs are also cheaper, though topping them up is painful so I’d personally stick with an eSIM there.
I expected to use an eSIM when I went abroad for a month last year. It turned out the providers offering "travel eSIM" are 2 to 4 times more expensive compared to buying a prepaid physical sim at the counter valid for 30 days.
Quick note that "the counter" may not exist or be hard/time consuming to track down and then there may be language barriers and also identity proof requirements that you can't meet. So service that's available and working as soon as your boots hit the ground do have some value.
I don't recall what it was called in the menu, but it was definitely possible to assume a struct on a particular address. Muscle memory tells me the button is U, even though actual memory fails me.
