I've dialed down a lot as well. The answers I got for my queries were too often plain wrong.
I instead started asking where I might look something up - in what man page, or in which documentation. Then I go read that.
This helps me build a better mental map about where information is found (e.g., in what man page), decreasing both my reliance on search engines, and LLMs in the long run.
LLMs have their uses, but they are just a tool, and an imprecise one at that.
> Some of you may wonder why ZubanLS isn’t open source, unlike Jedi. The honest answer is that open source never worked out for me financially. Beyond some small donations and a small recurring compensation from Tidelift, I was never able to make a living from it. I'm no longer a student, and with a family to support, it became clear I needed a sustainable path.
> Joining a company like Astral (makers of Ruff) could have been an option—but I’ve also grown skeptical of venture capital as a model. VC-backed companies often need to aim for massive success or risk disappearing within a decade. I want to build something that lasts. I’ve already spent more than ten years in this space, and I plan to continue.
Can anyone explain the `wrmsr -a 0xc0011029 $(($(rdmsr -c 0xc0011029) | (1<<9)))`? It seems to help on my system, but I don't understand what it does, and I don't know how to unset it.
As long as there is creativity in programming, and I think there is a fair bit of that, AI is just going to be a tool.
GPT-4 is great at sourcing human knowledge, but I think it can't really walk unbeaten paths. This is where humans shine.
Case in point: I tried to ask the AI to come up with a new World War Z chapter set in Switzerland, and it was unable to produce anything original. I had to keep feeding it ideas, so it could add something new and interesting.
This. I moved multiple web applications from bare-metal to the cloud, and many times had to refactor code to do hit the disk less. What worked before: Dozens of SQL queries, or loading session data from a file - for each request - it all was too slow in the cloud.
Also: I find that many engineers wouldn't be able to tell you if the storage attached to their VMs was networked, or actual hardware on the hypervisor. A distinction that has big implications for throughput and latency.
Same. Our compute hosts are generally not using 100% of their cores at all time.
There’s computes that are not full, computes that run rather diverse tenants, and even the fully utilized computes responsible for CPU optimized VMs have enough variance in their workload for frequency scaling to occur.
I don't think that's the case everywhere. In Switzerland at least, if I send you money by mistake, I can claim that money back and you would have to give it back to me, even if you have spent it already - unless you can reasonably claim that you were unaware that this money was not meant for you.
As a result, if I received a large amount of money that I coulnd't explain, I would contact the sender or the bank first. I can't just spend it immeditaly and claim that I had no idea.