Who needs utility or use cases now when you can jump right to supremacy in a market you make up the rules forl

I have started to notice a push to replace 'entry level' work with AI, and don't really seem to have a plan for how to train up new people or have them gain experience outside hiring already experienced people.

Though at my company, the big hope is that formalized 'design practices' will get around the training problem. No need to gain experience when there are hundreds of documents outlining the wisdom of random people, and a requirement to explain which ones you used and why you didn't use others for a project.

Seriously, they want us to justify things like 'used a return statement somewhere other than the last line in a function, in any language', and believe this will mean no longer needing to train people.

I think the key difference is that when we talk about revenge in most industries, it is, as you say, bombs or poisons or shootings etc. However in IT, what you generally see is a revenge that is an inconvenience and costs money, but that is about it. That element of 'the only thing hurt is the company's pocketbook' goes a long way to creating a romantic fantasy, esp since it i in response to the company hurting their pocketbook. Thus it regains a sense of balance instead of being disproportionate.

All that science standing in the way of our perfect future after all.

This is one of the problems with having a party that idealizes CEOs, part of that idealization means they also see private companies as the ideal form of organization, thus governments should operate under the same principles. Though most large companies also have healthy amounts of debt, so it also speaks their fantasy image of how private institutions actually work.

If a language can be designed to not allow something, and a compiler can be written to disallow something, that is within the purview of what static code analysis can handle. If you want the runtime component, that is just a library since all they are doing is swapping out the standard library, which again, you don't need a new language for, just pass an option to the already existing compiler.

I suspect that this is the US being the greater fool, letting the people who were unsure how they were going to convert their holdings to USD to start cashing out with the US buying up their worthless coins.

Because language design is fun, and the era of major languages springing from the ether is kinda over, so projects like this are people's rare opportunities to do language work.

You would think that improving static code analysis would be a better fit for the actual problems. No need to fork a language, if the compiler can be altered to detect such problems then a linter could too.

Our management is pushing AI _HARD_ right now. We work in aerospace, our stuff goes on aircraft and influences things like 'does everyone die when they use this?'. In the past, the FAA would step in and say 'no, this new fangled unpredictable tool is NOT appropriate for aircraft (seriously, we are not even allowed to use OOP because of how difficult it makes predicting and validating bytecode)'.. but now.. I suspect the 'new' FAA is going to be a lot more ok with it.

Something proponents tend to forget is that for a tool to replace another tool, it not only has to do as well, but substantially better, in order for users to actually adopt it. Which TBH, much of the industry has kinda given up on and has moved to forced upgrades instead. Notice that a lot of the push behind use of chatbots and other AI tools isn't coming from users, but managers and owners listening to other managers and owners then mandating what the people they have power over should be using. The main user group who seems to embrace them are the ones who don't actually know the existing tools yet or have the experience to know if something is working correctly or not.

Yeah, something I would really love to see are statistics on how many people keep using them. Right now they are being pushed HARD, with many companies (or at least executives who want to be seen as visionary) mandating their usage, and they are really popular with early career people who have not really built up much experience yet. The pattern I have seen though is people use them for a while, are really excited, and slowly the novelty wears off and the additional workload of actually making them work sets in, then people abandon them.

Right now my team is being tasked with evaluating a new 'this is the future' chatbot for our work. There are a few people who were part of the last evaluation and are already warning management that things will seem great through initial evaluation then go downhill when they are applied to actual tasks that need to be done correctly. But AI is the new secret sauce that is going to bring down their costs (or at minimal shift their costs from paying workers for their time to paying services for the bots).. so it is really hard to convince them of anything other than their own vision.

You cannot be conscious without daydreaming. The brain is perpetually recreating the past or projecting the future. Indeed, that is all it does, the present isn't important to it. There's no survival value in knowing about now, only in correlating with past threats/safety and determine what to do next.

As such, the brain is always jumping between past and future, perpetually daydreaming.

The Chinese Room wouldn't hold for any mind that used quantum effects either as described by Penrose or by Conway. Both of these, however, have/had no problem with the idea of a conscious computer, just not a Turing Machine class of computer.

Alan Turing was fundamentally a mathematician and a logician. From this standpoint, we can understand the Turing test to mean if f(x) lies consistently within the range of outputs of all possible g(x) in the set of conscious humans, then there is (obviously) no test you can perform to show f(x) isn't human.

In other words, it's not enough to appear human on a fairly consistent basis to one person. That's not the test. You have to define a valid range and prove that no output (without exception) will step outside that range.

The test, as written, is not the mathematical sense he would have been coming from. The mathematical sense is not a subjective freely one, but rather a rigorous validation that the system under observation is indistinguishable from what would constitute a valid member of the set.

This is not what Dawkin achieved.