jj4211

Yes, the "did I stutter?" If everything is a file philosophy.

Sadly, the browser variants lack functionality. In my experience, particularly powerpoint, a colleague says the same of excel.

Thankfully, I rarely need that functionality so I only have to resort to Windows once every few months.

Between libre and office in the browser, 99% of my needs are met.

Collaborating on a powerpoint however, it's very quickly 'powerpoint application only', whether I like it or not.

My company issued me a computer with Windows just fine. Except for Office 365, everything else I do is much better to have Linux native on it. So they are perfectly content to help me run Windows, I still don't want to run windows.

I was expecting you to say just use it in a browser, which actually I do 99.5% of the time. However, just last week I had to boot into Windows because of being forced to work a powerpoint presentation in ways the online powerpoint simply won't touch. Similarly, I've heard people speak the same of Excel, but I haven't experienced that one personally.

Incidentally, booting into that Windows install, the updates took about 6 hours before it was happy. Between Microsoft and my work's endpoint management, it's a nightmare to try to maintain the work Windows load. I had a similarly neglected Linux laptop and it took about 30 minutes to chew through an even larger window of updates.

Just saying that if I thought it could work, I might be inclined to use ASI as a bogeyman to induce people to actually do something to reign in this mess, even if I didn't believe it. Not saying that Bernie consciously employs this strategy, but it's a possibility.

Fair point about such a strategy potentially backfiring if it doesn't work fully, then the risk of 'the other guy' becomes a risk of making everything even worse.

Note that even if the ASI problem isn't realistically a threat, either way the AI companies are absolutely a serious problem, not merely potentially. For most folks, half their retirement is at this point at the whims of a handful of AI tech bros. Putting aside all arguments about ASI, that alone is something worthy of putting some control over them.

I can't say whether he has bought in or not, but this could be useful toward the end of curtailing some of the most damaging economic ambitions, so long as it is applied across the board without "trusted established players" somehow being given a pass.

So it could him buying it at face value, and frankly who can blame him. Trying to discern reality from hype is tricky in this scenario, especially for a politician.

However whether the scenario is substantitve or simply everyone is spending and building and planning as if it were substantive, his reaction is equally useful, as slowing our roll is important in either case.

His style of 'buy-in' wouldn't be as risky as the more pro-AI buy-in, like Trump.

"Our gods are dead. Ancient Klingon warriors slew them a millennia ago. They were... more trouble than they were worth."

The devnode that is used to reference the block is a file, so you come back to blocks just 'being offsets into some file'.

Though system calls, ioctls and networking credibly get away from file semantics

Missing stage 7: Not everything is a file.

Linux is closer than other *nixes (sysctl equivalents in some platforms are not files, but they are files in Linux).

However get very far into network and the file-like behaviors become more limited. From an application, at least you have a decent chance of having unseekable, but otherwise 'file-like' interaction with stream protocols. Then you need recvmsg/sendmsg. Then you want to deal with low level interaction with the network stack and dealing with NETLINK and such..

It's still consistent with my experience, but in certain contexts, can be corrected if the output can be processed in some context that can auto detect wrong output. For example if you have a unit test to pass, then it can be wrong several times but doesn't matter because it auto runs and gets directed to try again.

In contexts without the possibility to objectively measure correctness, it is still pretty bad.

Think there is a much higher immediate risk for new developers trying to get going in the market. So much management right now is pretty comfortable that new programmers with zero real world experience offers nothing the GenAI can't do under the supervision of a senior developer.

Problem is historically new developers weren't trusted to make any decisions if more senior employees were aware of such scenarios, so a new developer was pretty much verbatim implement guidance from more senior teams just deal with obvious rote tedium associated with getting it to code. Also the senior dev will have to fix all your screw ups when you inevitably botch it a few times. Awfully similar to the role GenAI can do now. So has to be some cultural shift to figure out how onboarding new generation developers looks.

Note I could be wrong about the middle paragraph, it's what the larger organization thinks, but I kind of got going by being 'senior' for lack of a company having options and moved from there. Similarly, I've never had success onboarding a junior that wouldn't take independent ownership. I'd get too frustrated at lack of growth and having to correct the same mistakes and be painfully specific, much like I do with GenAI. A new college hire that had his own vision and ability to take vague guidance and implement his vision, that I could work with, but it has happened once in a great while.