If the claims about AI capabilities are true, it's matter of time before all providers to switch to homegrown solutions.
“VMWare costs too much. Our use case is incredibly niche and nothing out there seems like it would fulfill our needs. Can you help me write a new stack using bash scripts and AWK?”
“What an exciting idea! Bash Scripting is a clever way to orchestrate 12,000 virtual machines. Here, let me provide you with a bunch of code to copy and paste without doing any requirements gathering whatsoever.
<pages of random bash scripts>
These bash scripts will let you run 12,000 virtual machines. Let me know if you need any explanation!”
“I forgot to me to mention we need to provision these using a rest api. Also these are globally distributed and require extensive monitoring”
“Not a problem! REST is an excellent way to communicate with the platform. I won’t bother asking any follow up questions or gathering any more requirements. Here is all the code you need:”
… and so on…
Just like with the image and video generators, I'm fishing to find evidence of machines building something useful - had very little success so far.
If the claims are true and these huge investments are not done in vain, not too far in the future we should be able to tell AI to build the tools we need, for example Microsoft should be expecting to discontinue MS Office either because someone can tell AI to make them an Excel instead of paying a subscription or tell AI to just do the job they will do in Excel.
It's very strange that they talk about replacing the developers with AI but somehow still sell software. Something doesn't add up.
There's a simple answer. This is the "replacement" fallacy that is all too common when hypothesizing the impact of tooling. Tooling increases the productivity of workers. This "replaces" workers in that you will need fewer of them to do the same job, but can't replace all of them because zero times anything is zero.
If you have 10 people shoveling, you can replace 9 workers of them with a backhoe. But a backhoe can't replace all 10 workers.
for example, salesforce proudly announced that it’s not hiring the developers this year because they are going to use AI. they don’t intend to make even greater software with the same developers using AI, they intend to replace them.
i’m getting the vibes that nobody at the top believes that AI actually can replace anybody. I find this interesting.
A box of literal rocks does not do anything on its own, even if we have tricked them into thinking even better than we did last year.
Literally for hundreds of years have people been fear mongering that some new technology will replace workers. It never happens. Technology displaces the available distribution of jobs, but it doesn't eliminate the need for people... and it never will, because human jobs are a human construct.
> salesforce proudly announced that it’s not hiring the developers this year because they are going to use AI.
Yeah, more bullshit. Translated, they really mean: "Hey our AI is so great that we don't have to hire developers anymore -- it's definitely not because we have 70k+ employees and our payroll is way too big -- it's because we're so good at AI and you should totally buy our software #EinsteinAIRocks #SalesforceOhana"
But it is genuinely replacing workers, in real terms. And more importantly it's replacing services, which in turn means jobs and revenue. So they too turn to electricity & automation, so they can do more... so they can remain competitive.
So uh yea, I 100% agree with you there, and wish that more people saw straight through it like you do.
Replacing whom? Can you name a single company that has laid off employees because AI is now doing the work?
Which is complete bullshit. Not only are they hiring right now, but having worked there, none of their engineering org's productivity pain points are going to be solved by AI.