I propose we use our powers of reasoning and knowledge of chemistry as a species to try to do better than just rolling the dice and hoping for the best. Economic suicide and blind luck are not the only options. But for some economic suicide appears to be the goal.
Yes, of course there are people who really want to do those. But op was presenting their main point of view related to "the slow evaporation" part.
Like op, more and more people can't get past the idea of starting or contributing to FOSS is no longer worth their time, due to various reasons, such as: their free labor of love was turned into another corporate money making tool.
OK? But I'm not arguing that there don't exist tensions between employers and employees, of course there are.
I'm saying that liking isolation and not wanting to interact with others isn't good for employers, and don't be surprised if they push back on it. It amounts to saying, I only want to do the fun parts of my job, not the boring but necessary bits, like co-ordinating with other humans.
Surely the C-suites of the company will employ whatever brilliant ideas that we talk over the water cooler. The water cooler is the fountain of innovation after all.
Perhaps they should put all the water coolers in the executives' offices, so they can listen to all the brilliant conversations that take place at the water cooler.
To be fair, when i worked at IBM Research (Watson), there were collaboration areas at the end of each hall.
They got used quite often, and there are plenty of times where someone noticed another team or person working on something and discovered it applied to what they were doing and collaborated.
One example from an area i know well - if you look at static single assignment form for compilers, which is the basis of all optimizing compilers these days, two people came up with the static single assignment part, but had no idea how to create it fast , and ran into some others whiteboarding control dependence for other reasons, and realized that it solved their problem.
Why is it that every time someone tried to push RTO agenda in a thread, "water cooler talk" has always been pushed like it is one of most important thing in a job?
Is "water cooler talk" the new "open office layout"?
> Is "water cooler talk" the new "open office layout"?
Yes. It used to be that "water cooler talk" was considered an unproductive waste of time. But for people pushing an RTO agenda, it suddenly became the go-to argument for RTO. Very often, the same PHBs use this to push RTO, who back in the day complained about people doing "water cooler talk" on company time.
At the same time, open office or cubicle layouts are proven to be detrimental to work output, concentration, health and happiness. But same as RTO, it is not about any of that, nor about collaboration. It is about PHBs fearing their loss of purpose and control.
> .. been pushed like it is one of most important thing in a job?
because it is actually one of the most important thing in a job. During Covid, work was boring as hell because there were no spontaneous interactions or small breaks to chit chat with other people.
And guess what, most of the colleagues I have talked to mentioned that they missed those kind of interactions. Barring some super introvert ones who just want to be left alone, but those are a tiny minority.
depends on the tool and purpose. There was skill in navigating a file system, and now the next generation (obbscured from folder systems by mobile) seem to be losing that ability.
You can look at it in two ways, neither are particularly wrong unless your job is in fact to navigate file systems.
Of course, LLMs would be more useful to many more people if they could be used without skill, and were as "useful" as a human mentor.
That's true, and they lack that capability. Many people seem to react as though this means they're missing all value, however. I find them incredibly useful; it just isn't possible to get much value out without investing effort myself.
Besides the ideas from istjohn, empath-nirvana, and rcarmo, you can also just "flip the script": solve for epsilon and report that as 1-delta confidence interval for the worst case data distribution as here: https://news.ycombinator.com/item?id=40388878
Best case error is of course zero, but if you look at my output then you will see as I did that the worst case is a very conservative bound (i.e. 15X bigger than what might "tend to happen". That matters a lot for "space usage" since the error =~ 1/sqrt(space) implying you need a lot more space for lower errors. 15^2 = 225X more space. Space optimization is usually well attended for this kind of problem. And, hey, maybe you know something about the input data distribution?
So, in addition to the worst case bound, average case errors under various distributional scenarios would be very interesting. Or even better "measuring as you go" enough distributional meta data to get a tighter error bound. That latter starts to sound like it's one of Knuth's Hard Questions Which if You Solve He'll Sign your PhD Thesis territory, though. Maybe a starting point would be some kind of online entropy(distribution) estimation, perhaps inspired by https://arxiv.org/abs/2105.07408 . And sure, maybe you need to bound the error ahead of time instead of inspecting it at any point in the stream.
You would want to calculate the threshold by choosing your target epsilon and delta and an 'm' equal to the largest conceivable size of the stream. Fortunately, the threshold increases with log(m), so it's inexpensive to anticipate several orders of magnitude more data than necessary. If you wanted, you could work backwards to calculate the actual 'epsilon' and 'delta' values for the actual 'm' of the stream after the fact.
You actually don't need to do that part in the algorithm. If you don't know the length of the list, you can just choose a threshold that seems reasonable and calculate the margin of error after you're done processing. (or i guess at whatever checkpoints you want if it's continuous)
In this example, they have the length of the list and choose the threshold to give them a desired margin of error.
