jillesvangurp

Skills that you need come and go. I owned a commodore 64 once and knew how to use it. All gone. I fired up an emulator a few years ago and then had a big "now what moment" where I realized I had no clue what to do with it. I had a vague memory of inserting tapes and issuing commands to load stuff from it. Function keys were important. And there was a manual I once had that explained it all. But no muscle memory. Nothing. I'm sure I could have figured it out eventually. Same with DOS. It's decades ago since I did anything on a DOS command line. All gone.

This doesn't matter. That stuff belongs in a museum. I've used AWS pretty extensively a few years ago. I switched to gcloud and other providers later. At some point in the last year, I started delegating all of the pesky work of interfacing with that stuff to AI. It's amazing. I've always referred to AWS documentation as a maze of misdirection. Just endless amounts of documentation mixed with marketing pages that are completely useless. All links lead back to those. So you are constantly fighting their marketing people to get at the bit you actually want which is a mess of fragmented reference documentation with parameters, endpoints, etc. Staying on top of that and keeping it in your head is a skill that will atrophy that I never cared about.

AI is great for figuring out micro detail like that. It's also great for coming up with a bigger plan that involves stringing together lots of tools, commands, scripts, and what not to execute a high level goal of "I want to run this thing over there". And with a bit of work you can actually do this the right way as well. It's fine. I never really liked devops to begin with. I can do it and I appreciate it done properly and can tell the difference. But I've always hated how it sucked up my time when what I really wanted to do was move our product forward. Instead of which I ended up dealing with a lot of arcane AWS or whatever crap that just needed dealing with. And there's always more of it. Just endless amount of micro detail. AI is great for getting on top of that.

The issue with LLMs is not that they write poorly but that we all use the same models and end up with the same non-unique voice. This can come over as a combination of bland, boring, and unauthentic. And it doesn't help that some LLM generated text reads like these LLMs are biased towards a style of communication which embraces a lot of hollow phrasing, disingenuous language, overly enthusiastic language, etc. No surprise, the key companies are based in California in an environment where people that work there probably communicate like that. For them this is authentic. But I'm a non native speaker in Berlin and if I send language like that to my colleagues, clients, or peers, I end up sounding like a jerk.

It's something that is of course fixable. Either with a lot of prompting and direction. Or by training on existing text that you wrote. To some extent some tools already do this. But of course if you only ever let AI generate what you say, how would you even develop your own voice?

On a meta point, I did write all of this myself.

Most countries in the world have to import their oil and gas. Which generally means EVs are very attractive as they can run on locally produced electricity. This is happening all over the world at this point. Including quite a few developing nations across Africa, Asia, South America, etc. Ethiopia banned imports of ICE vehicles a few years ago. They are going all in on EVs. Same for Kenya I believe. They aren't doing this to save the planet but because oil imports are expensive, especially if you have to pay for it in dollars and don't have a lot of exported goods to earn those dollars with like is true for a lot of developing nations.

The only countries trying to slow this down currently are those with a lot of vested interest in oil and gas. Or a lot of legacy car production. Everyone else is cutting their dependence on importing either of those things.

Blaming AI is correct but perhaps not the way you think. I think a lot of people actually use other people's AI tools to do their searches now. Google's attempt to grab that back via their own AI tools is so far not really working. Google is no longer the go to tool for answering questions or doing research.

I used perplexity for some time. These days, I have claude and chatgpt. They search on my behalf. Sometimes I still do a Google search but I mostly just use it if I want to open a website that I don't remember the url of.

> Sharing all my financial info with LLM companies is not something that I like.

The price of convenience and massive time savings is taking some risks on that front. We run a small business. We heavily depend on codex doing all sorts of administrative stuff that in the past either wasn't happening at all, wasn't happening properly, too late, or was being outsourced to people with various degrees of flakiness/incompetence/ineptness. We got rid of all of those and freed up a lot of our time.

As a small business you are always reacting to incoming demands on your time (accountants, taxes, invoices, insurances, etc.) for all sorts of reasons that boil down to having to do a lot of detective work in documents, filling in lots of forms, responding to queries, etc. AI tools can do essentially all of that now faster and better than most people can. Even the complicated stuff that involves fiddling with difficult tools, poorly designed websites, convoluted processes, etc.

Half the success here is getting organized on how and where you store data. A lot of SAAS tools actually make this harder, not easier. It really matters which bank you use, for example. Some banks have decent APIs and tools and export options. Others don't. Same with accountants. Some are more forward thinking than others.

It's a bit of a leap of faith when you start delegating important stuff to LLMs. And then you realize it's mostly fine and that it's doing a better job than you would. You still should double check the output. And you probably should still use an accountant to sign off on things and do some final checks. But don't fill in forms manually or go hunting through your drive for some document or information.

I wouldn't mind either. But I do see the point of different countries wanting to be in different time zones. I live in Berlin. My parents are in the Netherlands. We're in the same timezone but there's a good 30-40 minute difference between when we have our sunrise and sunset. Very noticeable when you do a video call around that time. I'll be sitting in the dark and it's still light on their side. The central European timezone spans from Spain in the west to Poland in the East. that's roughly 9 degrees west longitude to 24 east. A normal timezone should be about 15 degrees (360/24), not 33. That's more than two hours difference between when the sun sets/rises from east to west.

So, I can see that how some countries like Spain and Poland might want to make different choices when it comes to losing or keeping the hour permanently as Spain is basically in the wrong timezone. If you want to have noon occur when the sun is highest in the sky, you need to be close to 15 degrees east (roughly the Polish border with Germany).

Much of Spain is technically west of Greenwich (0 degrees longitude). Which means it should be GMT. You could make the same argument about France and the Netherlands being closer to Greenwich than the Polish border with Germany. I'm at 13 degrees east in Berlin, so pretty close.

Countries further north tend to care less about this because most of the year it's either too dark or too light most of the time. I used to live in Sweden and Finland. Daylight saving doesn't do anything useful there. There's not much to save when sunrise is just before noon in the winter and shortly after midnight in the summer. In Helsinki (not in the same timezone), the sun is already rising around 2am in the summer and it's still dark at 10am in the winter.

A simple observation here is that historically inter people communication is the main scaling bottleneck in software teams. It's what's behind the main premise in the mythical man month that adding more people to a team makes it slower. You increase the man months, but you are adding so much overhead that it's a net loss.

It's why big software companies struggle getting big things done whereas small and nimble startups tend to run circles around those. Google, Apple, Amazon, and Microsoft are great examples of companies that are really struggling with doing anything outside of the stuff they are already know how to do despite adding people by the thousands/tens of thousands. The more people they have the less likely any of them is going to be to do something completely new and different.

How does AI impact all this? It makes the problem (inter people communication) much worse. Imagine five people doing things manually. They have to sync up regularly. But there's only so much they can do in a day so they can stay in sync relatively easy. Scale that to 50. Now your meeting overhead becomes very non trivial.

Now add AI to the mix. Each of these people goes off and maxes out their tokens every day. They are refactoring, adding new sub systems/products, new UIs, etc. How do they coordinate their actions? How are they going to communicate their intentions and actions to each other. How do they brief each other on what's already been done?

The day to day scope of the massive amount of changes done by each of those individuals quickly grows way beyond whatever fits in typical/reasonable meeting structures. So, all the obvious things start happening: people either don't communicate or communicate poorly. They work against each other, break each other's work, etc. Everybody has a hard time staying on whatever is happening. The easy implication is exactly what we see happening in most big organizations: not much happening at all. They were already bottle necked on their size. AI made the problem much worse and as a result they aren't getting more done. At best they are maybe getting about the same stuff done with a bit less people.

AI clearly favors setups where you can minimize all of the communication. Small 1-3 person teams are probably a lot more optimal here. They always were and e.g. Jeff Bezos figured that out early in the life of Amazon (the two pizza rule). But AI just makes it much more obvious. Remove the need to sync lots of people and you free up time to actually do stuff (with or without AI).

The same dynamic probably holds to some extent with using multiple agents. If you are burning a lot of tokens and time on getting agents to sync with each other, that overhead is only going to get worse if you add more agents probably. And if you need to be in the loop to micro manage them, you become the bottleneck. Even if you just work by yourself. The trick is providing enough autonomy so you don't have to micro manage as much. And structuring your agents to minimize the amount of other agents each agent has to sync up with. The two pizza rule still holds.

I get a lot of those in Germany. The countries they originate from change. A few months ago it was random numbers in the US. Then it switched to the UK. The last few weeks it's from some random numbers in Germany. The numbers rotate and I never get more than one or two calls from the same number. I always block them but it seems these guys have an endless supply of random numbers to switch to. I had four of them today. I just bulk block them every few days. I don't think anyone cares.

No idea what kind of scams they are pushing but it seems very cheap and easy to abuse the system like this. I wish we could escalate this to operators and force them to deal with abusive customers properly like they clearly can't be bothered to do right now. Some countries probably have rules and regulations. Like Germany probably has. And yet, I get spam calls from German numbers on my own German number. So, clearly whatever there is, isn't working all that well.

The speed at which you can switch is probably proportional to your dependence. But the thing with insuring against risk is that you need to cover all of them and not just cherry pick them.

If you are running your own models on your own hardware you take a bunch of risk as well. Is your hardware capable enough to run your models? Do you have enough of it and access to more when you need to? And is the stuff that you picked still going to be good enough in six months? And can you trust those model providers that you picked to keep up with the rest of the industry or are they going to fall behind and leave you stuck with obsolete but still very expensive technology.

Of course, this is an article about IBM who are in the business of selling consulting and services at a premium. They would probably love companies to self host their models on infrastructure that IBM supplies and consults on (for obvious reasons). The question is, should you buy what they are selling or can you do better? My simple observation here is that if companies like IBM are getting excited it's probably because they are looking to separate a lot of companies from some pretty significant cash. So, I'd be wary to buy what they are selling. But that's just me.

In a rapidly evolving market with extreme levels of uncertainty around tech and what will happen next, being reluctant to get pinned down on the wrong solutions is probably a good instinct. The biggest risk here is making the wrong bets. And you need to look at what it's going to cost you if you get it wrong and what it's going to cost you to fix it.

I can only speculate. But my guess is that they don't anticipate additional mac sales if they were to ship this feature. So, the only reason to do this for them would be if it would bring in more revenue. And given the cost overhead related to patents that I outlined, it might just not be that interesting. My guess is that a lot of the agreements they have for this might cover iphones and ipads but not laptops.

It's not just Apple. 5G stacks in laptop are just very uncommon. Most vendors seem to independently come to the conclusion that they are better off not bundling that. Of course the difference is that Apple now has their own modems and chips for this. Which might change things but so far hasn't.

The reasons Apple want to vertically integrate everything and own it are easier to grasp: chips, firmware and software made by them are easier to control and upgrade by them. With qualcomm, they are basically getting them, their development teams, their supply chains for chips on the critical path to shipping products. And they are outsourcing what is a pretty major part of the whole phone experience to a third party.

Apple has been working on their own 5G modem for many years before they shipped it. I don't know for sure but I would not be surprised if efforts for that started when Steve Jobs was still alive even. This was always the logical move for them to make as soon as they got the idea of doing their own chipsets for their phones.

If you want to look at a place that is comparable to (parts of) the US that manages to do things about 3-4x cheaper, look at Australia. They seem to have stumbled on something that works ridiculously well. You can get a quote and stuff installed in something like a few weeks at most. It's done right and efficiently without a lot of bullshit.

Other places (like the EU and especially the US) have regulatory problems around anything to do with energy. This is mostly about ass coverage resulting in a lot of pressure for certification, related fees, grid operators insisting on a lot of overhead, building codes requiring a lot of faffing about for the business of wiring together what is essentially plug and play stuff, etc.

If you have an off grid shed, providing power to it is cheap and easy because you have none of that. You buy the hardware on amazon or wherever. It arrives, you wire it up. Job done. And if you are smart you might let a qualified electrician do it right for you or at least check your work (so your shed doesn't burn down and nobody dies from electrical shocks). But fundamentally it's not that hard and this stuff is mostly pretty safe and easy to work with. Especially compared to the things people burn down their sheds with traditionally (e.g. generators).

There's a lot of optimization potential here and most of that is completely non technical.

Usually solar and storage are a package deal these days for grid scale solutions. Mostly, storage is not really about providing long term storage but short term. It used to be that the sweet spot that makes the most economic sense was something like 2 or 4 hours of power (relative to the maximum output of the solar panels) but these days batteries are a lot cheaper and 8 hours or more is becoming fairly common. The point of storage is not really to cover for all the downtime when the sun doesn't shine but to give grid operators enough lead time to bring online extra generation or buy capacity on the open market and adjust their planning. Mostly a few hours is all they need to ensure the grid stays up and running through daily peak usage windows.

In aggregate, grid operators have a lot of other tools at their disposal via flexible pricing, flexible demand, and automated systems that can adapt to this to e.g. charge cars when there is a surplus of energy, turn on/off water boilers, etc. There's a lot of innovation around this topic in this space. This by the way is as much as making sure generated power doesn't go to waste than it is about ensuring there is enough power. They use batteries to time shift excess power to when they can utilize it.

Interconnected grids means that e.g. an operator in Denmark can import/export to and from the UK, Norway, Germany, etc. On a windy day they'll be exporting wind power. On a sunny day, they might add some solar to that mix. When there's no wind or sun in Denmark, they might pull in Scottish/German wind, Norwegian hydro, Swedish nuclear, etc. Batteries are being deployed all over the place at a pace that is only exceeded by the pace at which solar is being deployed. Whenever there's a surplus energy, batteries get charged up. It's very rare for all of the available generation to disappear all at the same time for long enough to drain all the batteries. And for those times of course fossil generation is still a thing as well, and will be for decades to come. There are a lot of gas based plants throughout Europe. As well as some nuclear plants of course. Singling out the EU market here but it's going to be similar elsewhere.

The point here is that there's no need to pair solar setups with days/weeks/months worth of storage. A few hours goes a long way.

Do you have a better reason? I think the price difference is probably too high to justify the feature. I don't see any technical reason why they wouldn't do this. But patent royalties and additional overhead related to device certification would definitely add some complexity and cost.

Debugging css issues related to print can be a bit of a PITA. Adding print options for dynamic content in a full stack page is not technically that hard. But getting it looking right can be a bit fussy. If you don't care about what things look like once printed, it's easy. But if you do care, it actually needs a bit of consideration.

I've done a few custom things with an application that generates labels for a label printer using just css and javascript. This is not that hard. But it does involve dealing with a lot of weirdness related to native print dialogs, vs. browser specific print dialogs (and their print previews), feature interaction with native printer settings and their scaling options, browsers adding footers to the printed document, and just general weirdness related to layout issues, page sizes, margins, etc.

My general feeling dealing with this is that CSS for print just gets a lot less scrutiny than regular CSS by browser developers.

One of the hacks you can do is generate a dom tree just for print. Replace the root of your dom tree with that, trigger the print, and then put the old dom tree back (with your full stack application and all the rest). Seems very hacky to have to do it like that. But it kind of works and is simple enough to do. And the benefit is that it stops the rest of your application (which probably does not take print into account at all) out of the equation and will stop that from interfering with your print layout. Conceptually, makes you wonder why you can't just hand a separate document tree + CSS to the print api call. This is a lot simpler than having one unified dom tree and one flattened unified CSS file. Separation of concerns is a good thing.

It's part of it. Another part of the reason is patent related royalties that they have to pay to the likes of Qualcomm, Nokia, etc. for any device that has a working 5G stack inside of it. Apple actually has had their own 5G modem implementation for a few years now; they clearly have the technology to deliver this with a clean full in house chip design.

But any device that has that embedded would have some royalty overhead added to the BOM. Which is why they've always had ipads with and without 5G and have so far never really bothered with adding this on laptops (which actually would make a lot of sense from a technical point of view). Operators would probably love that because it's a nice up-sell opportunity on monthly subscriptions. And with soft sims there would not be a need for a hardware sim card.

For the same reason, you don't see this feature on many other laptops either. There are a few obviously but it's a less common feature than you'd expect given how important good connectivity is to most laptop users.