I apologize for write "just" FTS. I know there is a lot of work involved and your article summed it up pretty damn well, in a way that makes it approachable for someone new to the topic.
I will keep a note of this article for next time I am asked about this topic.
My take is that they love apps for the simple reason they show as easily clickable buttons to enter. Most dread having to type a Web address for something they regularly do. Few know that they can create an "app-icon" from any Web page. Should be advertised more.
I had a new colleague on the team, who I had to on-board. I gave him a few simple tasks, just to get him into the whole setup. He literally copy/pasted my task description into Claude and asked it to complete the task. To begin with, I didn't suspect this, so when he asked for more help, I gladly wrote up a detailed explanation with more information and detail for him to learn. Little did I know, he never read it but put it directly into Claude. Not even sure how I should handle it, but my first instinct was to get extremely annoyed.
Maybe this is just how things are going to be. But in that case, I'm done spending my time being a helpful idiot talking indirectly to a robot through another person.
> Not even sure how I should handle it, but my first instinct was to get extremely annoyed.
My first instinct would be to have a /very/ direct conversation with that person and their manager, and the follow-on would be to escalate it further leading to their termination. That sort of behavior is unprofessional in the extreme, even in the era of AI.
"Demons are deceptive by nature, and typically speak with humans for a specific purpose, such as securing mercy or lowering vigilance. They treat language as a tool, using words without truly grasping their meaning. ..."
SQL, OLAP, Primary use case is fast aggregations on append only data, like usage analytics.
It's fast, it's........ really fast!!
But you need to get comfortable with their extended SQL dialect that forces you to think a little different than with usual SQL if you want to keep perf high.
This seems like the kind of things you can do before you get kids and have real responsibilities. Then you need to get back to reality. Sounds fun though and I would have liked to experience it.
It seems to me that it's based more on the use cases than the speed of improvements to the technology itself.
There were times when a CPU four times as fast changed what I could do with a PC. For a long time, with Moore's Law in full swing, we reliably got that sort of improvement every three years, and PCs older than that were widely seen as obsolete. Today that would only speed up batch jobs for me and have no impact on any of my workflows.
Some sort of on-device AI thing is probably the next threshold for PCs. I don't think there's anything production-ready and compelling right now, but I can imagine useful automation features when it gets good enough.
It's also that software back in the day was much more fine-tuned to use the very limited resources as well as possible, so getting a better CPU would visibly speed up things.
Somewhere in the late 2000s, the CPUs got powerful and cheap enough (in the sense of "cents per MHz") that it shifted from having to be creative to get your programs to perform at acceptable speed, to not having to and instead focusing on delivering marketable software faster.
The only thing nowadays I can imagine requiring a substantial amount of raw processing power would be on-device AI processing, but that doesn't seem to be the case here, as large parts of the processing is still done in the cloud.
Editing high-resolution video comes to mind as an example of something that will still stress most PCs. Being able to scrub through footage and preview the effect of edits in real time without using low-res proxy files demands a lot of performance. I think higher-spec modern PCs are there for 4K, but I'm not sure about higher.
Great read, although a shame that it didn't go any further than adding the write cache SSD solution, which must have been many years ago. I was hoping for a little more recent info on the EBS architecture.
Blocking the final result is no longer a real block, when all it requires is a vague prompt to replicate it.
We've entered a strange territory.
reply