hibikir

There are things one can call negative that are really realism trying to save you from very bad decisions. See the negativity of telling the boss at Theranos that the tests don't work. I've seen many a startup fail in large part because they spent their runway unwisely, as cheap tests that showed that the existing direction had structural problems were just seen as negativity speaking.

There is such thing as negativity that will ruin your mental health, but the other side is just embracing delusion. We can see examples of successful people that keep embracing the delusion and succeed. But there are many that end up in unrecoverable positions after they burn themselves to the ground in said delusion. Around me, there's far more of the latter than the former.

And yet one can argue that the "garbage businesses" you label are the ones doing their best at matching market demand. See how well things are going for other entrants, like Google, Microsoft or Meta. Putting Copilot in every app makes a whole lot less sense than Cowork, Codex and the like. They just are doing worse things while attached to traditional money fountains, like ads in social media and search.

In some ways it is crazy, but at the same time, anyone that has seen typical efforts of "digital transformation" in aging Fortune 500 companies is going to be skeptical of said company's abilities to become relatively efficient. I always considered that new entrants that don't think like they do would eventually win once the value of putting tech in the middle of decision making, getting rid of large divisions worth of cruft was worth enough. Trying to do the very same thing by competing with everyone is similar once you squint. You get extra money from your parent company to throw around during growth or massive reorg, but still have the problem of iffy incentives we currently see in the modern tech holding company, where what is good for the company, for consumers, and for the random VP have nothing to do with each other.

Unsurprising to those of us that has been screaming from the rooftops that human code reviews were barely catching anything in your typical corporate environment. The majority of the value was pointing out missing tests, because breaking bugs got past human eyeballs on a regular basis, even in very good companies.

It's not that the LLM code is that great, but that many overestimated the human code and the human review.

Depends on the layers you are working on. Security is top priority on a lot of things that are visible to the internet, but in a lot of large companies, the footprint of the whole org is so much larger that than, and there many companies are deferring many security issues. So then you see a lot of speed there, when the consumer accessible systems are the focus of the security push.

And you still have to make sure all the real constraints are in. I had to fix a proven correct distributed algorithm that just happened to believe updates were done atomically and with no latency.

Except when there's an actual urgent patch: See recent PHP or Ruby ecosystem updates. Then you are just a sitting duck for a week

Shared code repositories for OSS are much older than npm. See at least maven central for Java.

What makes npm a more relevant target is a much smaller standard library and thinner dependencies, leading to more chances that a random project will have code of suspicious provenance. The vast majority of what a backend Java project depends on was probably written either by the same 3 companies, or large, established OSS repos. So this is relatively solvable if people in steering committees decided they cared. It's improving, but I'd argue not fast enough.

It depends on the political system. In many a parliamentary system, voting differently than what you are told means you will not be in the party list ever again, so for the day to day, there's very little convincing effort when you have the majority. And especially in systems where there is very little legal money in politics, the carrot is just ability to get access for not-so-legal money. So party leadership is a matter of having enough businessmen friends, and having will to power in a rather unclean word.

The whole "you are just one vote" doesn't really apply there. As the 30th legislator on your party list, you show up and vote as you are told. And in systems like the one in the US, barring rare moments, it's also not all that different. The card that your party will abandon you, and that if you want a position of any importance you have to play ball means many a junior congressperson just votes as they are told, and spends time fundraising.

While there are incentive problems across all levels, The political quagmires that make it impossible to move at any reasonable speed, and that lead to counterproductive ideas getting large budgets are in my experience very common at the highest sizes. It's true that on the other edge of the spectrum, small startups make poor choices, but many of said startups will fail due to said choices. So by the time you have a viable company making good money, the really badly aligned companies are gone.

So when I look throughout my long employment career for best security and organizational practices, all the highlights are still relatively, relatively undersized companies, with headcounts in the 3-4 figures. It's as they grow, and start to be multi-headed behemoths, that they decay again, as what is good for the responsable mid-level exec, for the business unit, and for the company as a whole start to have nothing to do with each other. That's when you can smell the decay in the hallways. A few large companies manage to stave off this rot a little more than others, but it's inevitable. Then nothing important can be done in a timely fashion.

Because for the same behavior, risks are going to be very different on the other side. In my family, for instance, we need significant interventions to get a reasonable risk due to all the bad genetic factors. In my wife's family, not so much. So knowing how serious the risk seems much better than just handing statins to absolutely everyone.

In some ways, US healthcare is already very aggressive for some forms of preventive medicine. Compare the colon cancer protocols across the world. Basically nobody else is recommending colonoscopies to the entire population with no cheaper tests, but that's the standard of care in the US. It's not free to receive, and it's expensive in general, but for some reason in the US it is standard, instead of starting with non-invasive tests.

And yes, regarding smoking, the US has had one of the most effective preventive actions in the world. Compare actual numbers of smokers in the US vs, say, Spain, France or Turkey. So it's not anywhere near as clear cut as what you are trying to make it sound.

Forget financial benefits: There's just a high risk of ending up in a crippling financial hole. Many schools have tuitions over 60k/yr. Add mandatory campus living costing 20-30k on top.

I have a child graduating high school right now. As a developer in the US, this means no child of mine is getting need based scholarships. So the question becomes whether it really makes sense to go to any of said expensive colleges vs handing the child 200k+ and going to a worse school where he gets merit scholarships worth 6 figures over the course of his stay.

When you put those kinds of prices out there, do not be surprised if people start to think it's all a ripoff.

And if you ingest patches in seconds, you now have an issue of checking that the patch is not poisonous. Many a modern outage comes from changes deployed too fast. Damned if you do, damned if you don't.

We already have as much parallelization as companies can afford: Hundreds, if not thousands of new corn varieties, coming from one lab. The budget is the limit. Testing is just expensive.

Process improvements come from failing faster, so you spend less money on your test. Finding the actual answer without testing would also be wonderful, but LLMs might be fast, but they aren't magic. Something like Alphafold saves time in some steps, but it's not just "hand me a candidate", but find completely novel verification mechanisms, which would be just as useful for humans. Either way, parallelization is unlikely to increase.

It's often not necessarily about visibility, but real iteration time.

For instance, let's say we are doing biology on plants, which is still much easier than doing human trials. No matter how motivated claude might get, you aren't getting a corn plant to grow in 40 seconds. Real world research pipelines involve stages where even greenhouses don't do, and sacks of grain are shipped to Hawaii, or Brazil, or wherever, to try to get practical data for a given test.

If your computer simulation is so accurate you can skip a step now done with biology, great, but it gets harder the larger your problem is. I might see if a molecule kills bacteria kind of quickly, but what does a molecule do to a real human being? Not so easy.

What can make sense, and what has been my personal experience, is that I have shifted my most draining work to the AI, so I can actually get more useful work done without getting exhausted. Just like 10 hours mining sitting inside of equipment is better than 10 hours with a pickaxe and a shovel, even if we ignore productivity

If I have a team of 20, making a lot of money, but I think I can do the exact same job with 15 because of AI productivity gains, should I keep the other 5 employed for the vibes?

The total net income doesn't matter: The move is management saying they are overstaffed for the work they think they need.

Assume Microsoft is expecting just a third of people to take the offer. It's just too many people to want to move to a different division that needs more labor: It's basically entire divisions that they'd have to create from whole cloth.

For most Microsoft employees, the maximum unemployment benefits in most states are a rounding error vs the buyout offer. Here in Missouri it's $320, maxed to 20 weeks. So for most local employees that might get this letter, It's less than a month of salary, and that without accounting for months of health benefits, which I bet the buyout includes.

Interesting perspective. As someone who didn't even grow up in the US, I saw Sim City as a builder of relatively low density, hard zoning cities that had little to do with the high density I grew up in. The idea of commercial zones didn't even make sense to me, as in Spain, such thing doesn't exist. Need for traffic to get to the commercial zones made little sense.

Solar seems to be working just fine in Spain, a place where almost nobody lives in a stand alone house, or even a row house. You still put solar in the building, but then you have industrial solar in places wi5 little agricultural value, and wind turbines installed on top of hills. All of this could work just fine in any sim city I remember, as you are just dedicating space to power, which is fine.

The world has more than one country, and most have basically no access to oil. They will go deep into solar first. See, for instance, how the mix between the Venezuelan takeover early in the year, plus the Hormuz situation, put Cuba's energy access into jeopardy, and how that meant moving heavily into solar, as at least that remains available once purchased.

So even if a place like the US goes lighter on solar that is reasonable, Places like Spain, Italy or Morocco will go all in just to minimize dependence on a resource a third party can control. It's the ultimate counter to the Hormuz situation. Squeeze access to oil for long enough, and you accelerate the move away from it.

If the rocket company was in the stock market as its own thing, not as an extremely complicated bundle, it would not be priced anywhere as high as it is now. I would expect that company to be healthy for quite a while too.

The actual difficulty we have with the modern tech company, which is at its worst with Tesla and SpaceX, is that we are really seeing a price or large conglomerates, so they are extremely hard to price. If they were all split into individual companies, how much of the market cap would go to each? I find the bundles to be rather anti-investor.

And no, it's not just those, and not really just tech companies, but that's where the problem is at its worst, as that's where expensive investments some of us consider worthless are tied with business worth something. Take Amazon. You are buying AWS, a logistics company, an ad company, and as part of this, there was always a devices division selling Alexa. I found Alexa to be a real loser from very early, but there was no way for investors to say what they thought the division was worth. The Amazon es execs could at any time decide to invest billions, or invest zero, and given the size of the conglomerate, the price change was going to not be all that high.

SpaceX has many components I would price near zero, where a lot of money is being spent. But as a pile of risky bets, the fact that they are a pile means it's harder to bet against any one, or ten of them. And if it merges with Tesla, that makes it worse. The difficulty in pricing the bundle makes it really easy to hide big mistakes for a while, until they are so big it all crashes. I don't think that is good for investors, or for having accurate pricing, which is one of the goals of a stock market.

Any corrupt organization anywhere will do its very best to protect itself from reputational loss whenever possible. It just happens that in some societies, the amount of protection those organizations can provide to themselves without quick consequences is much higher than in others.

In the US, a lot of supposed government organizations are basically immune to pressure from elected officials. A police department can do more to harm the standing of a mayor than the mayor can do to said department. In fact, sometimes you even have state involvement to make sure the mayor looks bad. See, for instance, the wonder of the St Louis Police Department. The state votes in a different way than the city, and the state passed laws to take over the local police department. Thus they control budgets and leadership, even though the actual taxpayers that foot the bill are local. As you might imagine, this leads to very strange incentives.

If the value of higher education is purely book learning, sure. But if the value is signaling and the creation of ingroups that will become future elites, they can be valuable even if you do not learn a thing in class. It's a debate about elite higher education that has been going on for decades, as top institutions refuse to grow their class sizes, even as they could find 4x the students that would fit qualifications.

Remember when we were told only an expert should install Linux, and everyone else should just ask an enthusiast for help? Well, now we all have an expert at hand. Might not be a world expert, up to date in the very latest changes, but more than enough to make a lot of things that seemed impossible doable.

You are talking about someone that spoke with a rather manipulated voice almost 24/7 for years. Of course she can act all day, every day. It's a big part of what her job was.

SBT cleared a lot of footguns in the last decade or so, but at the time where Scala adoption was popular, it still had a lot of weird syntax, including the strange, meaningful whitespace, and enough DSL syntax to scare people.

Getting it all to work nicely with easy syntax without breaking backwards compatibility at all is not easy. It's the same reason Option types don't take off in Java, as the effort to make them work in an environment with no features for null safety is going to break things.

It's a very conservative language that wants to minimize breakage, and the tradeoff is not without costs. I would argue than in LLM-world, the right tradeoffs change, but I don't think a language like Java is going to assume everyone has an LLM to do migrations for free any time soon.

80% of Java features in the last few years are Scala features with worse syntax, because the better syntax would not be backwards compatible. It's not good for Scala 3 adoption in industry , but I would argue Martin has no interest whatsoever helping that. He's happy with his research language that somehow has enough good features some companies put it in prod. Languages that care about industrial usage just organize themselves differently.

Don't forget that the Imperium runs on hypocrisy. Large percentages of the Adeptus Mechanicus is breaking their own rules, including working with abominable intelligences. In fact, what they are dealing with in their main factory world in Mars is a supposedly secret abominable intelligence. And it is, in fact, a shard of an alien god that turned other races into synthetic robots. Banned in public, the center of their cult in reality, even if many probably don't understand what in the world they are doing.

So yes, anything that is supposed policy in the Imperium is supposed to be taken with buckets of salt.