Data and Analytics Business Ideas

Lets crash and burn, baby If you’ve been working with swe for some time and give a shit about the actual work you put out maybe you agree that the code being generated by AI is slop. And it has always been since the beginning, although it got “better” over time. Most of us now basically pull the prompt lever of the AI slot machine, hope for the best and **IF** you actually give a minimum shit about what you’re doing (and the project you are working on) you have to go through all the slop, the overengineering, the bullshit try/catch/except, the absolute ball of mud of an architecture and the trash that comes out of AI to make it (hopefully) “good enough”. And thats not an easy task, the cognitive load is huge and **GOD FORBID** you actually write your own code like you did before all this. All of this justified by the HUGE gain in productivity of around 20% (given some recent studies, which I have many questions about).  Now you add to the equation the AI labs CEOs shouting to the world that coding and, hell, even swe as a whole are “solved” (whatever the fuck that means). (Yeah, good job Boris, 500k of LOC for the stupidest most basic CLI plagued by all kinds of bugs that were not solved for sooo long - cant copy from terminal, SCROLL for christ sakes). Then (what seems to be) all managers, C-levels, directors who haven't touched a single line of code all eat that shit like its the human centipede movie. Productivity is now expected to be 10000% higher, since a feature is only one prompt away now in their mind. And YOU (if you give a shit) try to argue that MAYBE the AI is not better than you. That MAYBE the AI bros that say that everything can be vibe coded now dont actually know what they are talking about, they are just replicating something they heard. But nooooo, now YOU are the lazy, deprecated, slow gear of the company. Hell, maybe the whole development team is a burden.  Since from my experience arguing and defending my work is nearly impossible because: ...

You can't have homogeneous data set. You can't have homogeneous data set. You can't have low-quality data. You have really low-quality data. You have really low-quality data. You have really high-quality data that can effectively high-quality data that can effectively high-quality data that can effectively train train train train these models and be effective. The train these models and be effective. The train these models and be effective. The second point is that this data set also second point is that this data set also second point is that this data set also allows you to create the best allows you to create the best allows you to create the best foundational models foundational models foundational models possible for the actual for the possible for the actual for the possible for the actual for the development of those algorithms.

- Drive implementation of planning, reporting, and analytics solutions, including - Develop and maintain financial models, dashboards, and management reporting that provide actionable insights into business performance, drivers, risks, and opportunities.