True that, but market will decide, if it will be too expensive the end user will not buy it.
We aren't in a truly free market that would allow for that. Keep in mind, there is quite the noisy crowd that doesn't want "AI" embedded into their
computer OS and yet it is still being done so (mainly because corpos are a bigger consumer demographic compared to us peons). Now, those that do, don't care, because they want it, but for everyone it is hoisted upon them and where there are few competitors in a market, regardless if they are expensive or not, if it is a staple good, people are going to have to figure out a way to get.
True that also. But all businesses tries to get a competitive advantage in the their market, Each year a company plans to increase their sales (at least it plans
). Ways to increase businesses are unlimited, each business will choose its path, some will win some will not, some will choose automation/AI as a solution to grow, some of them will win some not. I think that our industry is not standardized and that's why one software will not be suitable for all
sign makers. Also A software has limited functions and can not incorporate all scenarios from all
sign makers way of doing things.
I don't disagree that that every company will try to get some advantage over the others, be it knowledge, automation or just sheer scale of operation. However, we are at a stage that of hype and everyone is trying one that to achieve this and that one thing isn't quite panning out, the fact that companies that are vendors of this tech are requiring their worker bees to use said tech as part of their eval, that should be telling. They, I would speculate, have unfettered access to this tech and yet still needing it to be required for their work? If it's good, people will use it (for some people, as bad as it still is, it's better compared to what they would produced and that's the more dangerous crowd actually), it wouldn't be needed for work eval.
Even in "standardized" industries, there is still going to be some uniqueness that people would benefit from some bespoke tool. There is a problem with having "AI" generate it (and this problem exist with people hand writing it as well, just from a different way) and any scaled tooling, it's not going to be easy to maintain. It will be a "trust me bro" situation, but the "bro" will be "AI". Code review is mind numbing as it is, it's even harder when some chatbot can spit out "code" at a much larger daily rate compared to a human, that's a lot of code to review that wasn't written by the person (that isn't easy, not even easy if a human is reading human written code as well or if it's the same person just with a time delay from last looking at it, there is some lag there, especially if how they do things has changed).
All of my internal tooling has been handwritten. Some rough around the edges and unpolished, I used whatever got something out working faster first and when time allowed refined it. Here is the thing, if some dependency breaks or if some API changes, I can easily go in there and fix it far quicker compared to trying to worm around some "AI" generated code that has a lot of verbosity to it.
Yes is true, but that is happening because wasn't trained good enough. Whatever you train it will do it, the outcome is as good as the input. Garbage in Garbage out
This very training itself, regardless if it was trained good (it wasn't) or not, the fact that it was trained and how it renders it's results is how this is very much not "AI". However, any commercial, ready to go "AI" that most people would pay for that isn't local is going to have whatever bent to it's learning that it's vendor wants (already have seen it and there have been some kerfuffle about it as well).
Keep in mind, I find it hard to believe that that internet hasn't already been scrapped by these bots and what any new data that they are getting is now the "slop", so we have an ouroboros situation going on right now.
This isn't even getting into the legal implications of the training data (be it for code or for art). Although that will more likely change for the corpos benefit since they have the lobby power to do something about that.
True, at leas the customer is warmed and has a direction that is easier for us to replicate, than to interview the customer to dig out what he wants!
Not necessary. Most customers don't really care about warnings and think of them as just "suggestions". I have yet to see a direction that is easy to replicate, typically because most tend not to know what is or isn't doable in the realm of physical construction. Or on the off chance that it is doable, the expense is such a sticker shock.
I wait for that day! If will ever come!
Not necessarily. Typically what happens is that current methods become so expensive that it isn't necessarily a breakthrough in tech, but some other previous expensive tech that now is starting to get "cheaper" by future of whatever was used before got so expensive. The change from whale fat to kerosene for example.
