williamsmj

> The speed limit is absolutely not the right single value in heavy rain / fog / snow / etc.

Obviously. The number I suggested is 1) an upper limit not a lower limit 2) a limit, not a suggestion 3) lower than the current upper limit on pretty much all cars.

> The speed limit is also not the single correct value in case of some freak occurrence where it's better for everyone involved to exceed the limit briefly to avoid collision.

Sure. That's why I suggested a number higher than the highest speed limit in North America.

I don't know if you're being obtuse or if I'm being unclear. I know what a legal speed limit is. I don't think people should drive at 100 mph in fog or near a school. (Did you think I was saying 100 mph should be the lower limit?!)

Apparently, vehicles are currently limited by the manufacturer to something like "whatever is safe for the engine" rather than "whatever the engine briefly capable of". Sure. Didn't know that. Thanks for telling me that. Not relevant to my point, but interesting.

I'm suggesting, in the absence of technology that allows us to limit car speeds dynamically based on local conditions and laws, that vehicles driven on public roads should be limited to min(100 mph, "whatever they are currently limited to, which is an insane speed no one has any business driving at on public roads").

I chose 100 mph because it's a bit higher than the highest legal speed limit on the continent where I live. I'd be fine with a smaller number though.

Well, there is a single value. It's called the speed limit. All roads in my country have one.

I trust you will support dynamic vehicle speed limiters based on the local limit when the technology catches up in a few years.

In the meantime, how about 100 mph (which is 15 mph faster than the highest speed limit in the continent of North America)?

OK. Sure would be cool if they set them to a number that reflected the fact that there are other road users.

Cars don't have speed limiters because otherwise their drivers wouldn't be able to get up to 130 mph on LA streets: https://news.yahoo.com/nurse-reached-speeds-130mph-crash-141....

For the same reason, cars need drift sticks otherwise takeovers would be less exciting: https://www.latimes.com/california/story/2022-08-22/street-t....

Re: test environments, of course Twitter has test environments, but if you're asking does Twitter have "staging.twitter.com" that matches the load, configuration and topolgy of prod, the answer is probably no because, like many companies at this scale they choose to handle moving stuff from dev to production with feature flags and incremental rollout techniques instead of the dev/staging/prod promotion process that is conventional wisdom at smaller scales. This choice increases velocity (especially if you have the platform team to support incremental rollouts and rollbacks) and building (or simulating) test environments that match prod would be insanely expensive, if it were even possible.

Like monorepos, this approach to ops might seem strange if you've only worked at small companies, and even then it's something reasonable people can disagree about. But I'm not convinced it's a useful discussion to have with an 88 year old senator. And it's not a sign that a company doesn't take engineering or security seriously, or that it has no idea what it's doing.

Re: resiliency, one of Twitter's datacenters is offline right now and twitter.com is up (https://www.cnn.com/2022/09/12/tech/twitter-data-center-cali.... So clearly the "small number" mudge is referring to is greater than one. Beyond that, yes, I'm absolutely sure that the loss of a "small number" of datacenters could take twitter offline and even result in permanant data loss (the data is in datacenters too!), because that statement is true of literally all tech companies. Twitter's expected uptime is a bunch of 9s. It's not 100.0%. The fact that it's not 100% is not evidence it "simply doesn't have a handle on its own infra any more".

And FWIW, resiliency appears to be pretty high for Twitter relative to the rest of the industry. They were down for 40 minutes early this summer, and haven't had an outage > 1 hour for 6 years (https://www.theguardian.com/technology/2022/jul/14/twitter-e...). Facebook (which I think most of us think of as pretty good at engineering, relative to the clown show most people seem to think Twitter is) was offline for 5.5 hours last year (https://en.wikipedia.org/wiki/2021_Facebook_outage).

Discussion is happening here https://news.ycombinator.com/item?id=32098112

> Generally, people who get PIP'd don't get severance, even if fired.

Not quite true at Amazon, the most infamous PIP user. You get put on "Focus". If you fail that, you are offered severance or to go on a PIP. True, once you're on the PIP, you're not going to get a severance offer. (This is all according to https://newsletter.pragmaticengineer.com/p/amazon, which is excerpted by the author here https://twitter.com/GergelyOrosz/status/1537810787473006593.)

How is this different from https://tradewater.us/?

I don't know if the "split" is justified, but "Some people wanted vimscript+lua instead of vimscript" is not an accurate description of the reasons for the existence of the neovim project. The goals and non-goals are stated clearly in the project charter: https://neovim.io/charter/.

"Research and development" is engineering.

Calculus and analysis is commonly the entry point to "serious" mathematics for engineers and physicists but it is certainly not the "only true way". It's not even particularly the usual way.

I recommend the content covered in MIT 6042J, Mathematics for Computer Science (https://ocw.mit.edu/courses/6-042j-mathematics-for-computer-...). The most important part of this course is logic and proof-writing, both of which are foundational tools of serious mathematics, that sometimes get elided if you come in through the physics side (i.e. via calculus and analysis). You will also learn discrete mathematics, i.e. "a bunch of stuff that is useful for cryptography". But I think the foundational stuff about logic and proofs is even more useful. It will rewire your brain in a way that prepares you to go deeper in maths.

The NLRA says "Nothing herein shall prohibit any individual employed as a supervisor from becoming or remaining a member of a labor organization, but no employer subject to this Act [subchapter] shall be compelled to deem individuals defined herein as supervisors as employees for the purpose of any law, either national or local, relating to collective bargaining."

...which effectively states that if a supervisor joins a union it's a no-op. The employer is not required to acknowledge their membership. In practice, unions specifically exclude managers for a bunch of obvious reasons related to their ability to bargain effectively on behalf of their members.

Your manager likely joined a union-ish entity open to supervisors. IIUC such entities are not protected by the NLRA and as such have few of the legal powers and protections that make a union a union. They’re basically affinity groups.

I agree with everything you say, but Goodfellow was a manager. Managers can’t join unions in the United States.

The article literally has a section "Lessons learned".

It doesn't suggest that you type a any old "specific thing". It specifically suggests that you type something to demonstrate you acknowledge the scale of the destructive operation you are about to perform.

The OP's main suggestion is along exactly those lines. It's actually less extreme, in some ways (see OP's "Lesson 1").

https://rachelbythebay.com/w/2020/10/26/num/ makes a similar point. I don't remember us saying people only have themselves to blame when that article was posted (https://news.ycombinator.com/item?id=24904204). Not sure why we're doing it on this post, which makes a number of completely reasonable and specific suggestions for improvement.

That makes sense if you assume Musk has learned something shocking about Twitter from the inside in the past couple of weeks. How likely is that?

It seems spectacularly unlikely to me that a demonstrably impatient person with a pre-existing thesis about Twitter would take the time to investigate and learn new things about its product/financials/culture/governance.

It also seems unlikely that a person with a known strong stomach for legal and ethical "flexibility" would be especially bothered by anything they did learn, assuming they took the trouble. But YMMV.

I agree it's bonkers that this is possible. I agree, in an environment where the SEC seems unable to stop him, Elon Musk might well consider 22% return worth doing. None of that is relevant to the point I'm making.

I'm contending the claim that the implicit (and completely pro forma!) threat in his SEC filing places "insane pressure" on the Twitter board. The the possibility of a 20+% stock price change will not play a large part in their calculations. It's just not a big deal in the context of such a volatile stock.

He demonstrated precisely how much pull he had when he bought 9%. It's about $10.

I think you should look at the historical volatility of TWTR before calling a 22% move "insane pressure".

It's also worth noting that the stock price is almost unchanged today as of this writing, which tells you that the market thinks he's not serious, or his lowball offer is going to be laughed out by the board, which significantly reduces the amount of "insane pressure" he's exerting.

This is standard operating procedure for takeovers, but I think you're massively overestimating the amount of pressure someone selling 9% of the company places on the stock price. Someone buying 9% (with an expecation that they would buy more) moved the stock $10. The same person selling 9% stock (with no possibility that can sell more) is unlikely to move the stock more than that.

No. Twitter absolutely has multiple repos. Some of them are very big, but they certainly have more than one. Why they have > 1 repo is a long story, but the post you're responding to is in the ballpark.

Here to recommend Sue Burke's Semiosis/Interference sci-fi duology about first contact with a intelligent plant life.

> Following this advice and using a tone that even resembles the tone of the comment or the original post will make you look/sound crazy. School boards (government boards in general) are used to seeing vocal, crazy people and are good at not taking crazy people seriously.

I agree with the general point of your comment but this is simply not true any more. I’m not sure it was ever true, but recent coverage of school boards and environmental review public consultation throughout the United States demonstrates that it certainly isn’t true in 2022.

See also http://www.ismytrainfucked.com/ for NYC subway updates.

If your implication here is that the very public backwards compatibility statement, the mailing lists where PEPs are discussed and approved/rejected, and the website python.org are a "locked filing cabinet", it may interest you to learn that this exact PEP was discussed over two threads and nearly 1000 posts on Hacker News: https://news.ycombinator.com/item?id=19948642, https://news.ycombinator.com/item?id=19985802.

Agreed.

I also want to highlight these packages will be part of the stdlib for python 3.12, which will EOL in October 2028. In other words, if this change will affect you then you have up to 6.5 years to figure out and implement a plan.

This exact possibility was discussed and decided against back in 2015. See https://lwn.net/Articles/640838/.

The tl;dr is that if you move a package into the standard library then you are restricted by the python feature release schedule (was every 18 months, now annual) and backwards compatibility requirements (see https://peps.python.org/pep-0387/).