Hacker Newsnew | past | comments | ask | show | jobs | submit | adamdegas's commentslogin

Wouldn't this same effect also lead to more potential hurricanes in the Isles as oceans continue to warm? I'm thinking something like Acapulco where Hurricane Otis rapidly progressed from a mild storm to a Cat 5 hurricane due to warm waters.

Strangely, though, the UK hasn't had too many hurricanes in its history, which is why I'm curious.

Edit: I may have answered my own question. Even with a higher likelihood of storms, I think the mountainous and hilly topography makes it hard for storms to really hit the UK, which might be why there have been few hurricanes in the past.


> the UK hasn't had too many hurricanes in its history

In other words: In Hertford, Hereford and Hampshire, hurricanes hardly ever happen.


Hurricanes need warm water to form and sustain themselves - not just relatively warm but absolutely warm. Like, 26 C. Minimum. Which doesn’t happen in the waters off the UK, Ireland, and France.


That's too fancy. I think you mean <center>...</center>


> vertically center


Battery life


Not so much; running Chrome headlessly can yield almost the same power usage as a graphical spawn one.


You're not supposed to run it on the same device. You're supposed to run it from a server and SSH into it to save power on the client device.


You would run browsh on a power unconstrained server that you SSH to from your power-constrained device.


The layout and font gives off a more early 2000s vibe than a 90s vibe for sure, or, more accurately, "early 2000s with 90s leftovers." The average Neocities page uses more serif fonts, basic table layouts. Also computers in the 90s weren't going to be able to handle that rain effect, smooth scrolling marquees.


It absolutely is a blend of modern and early web.


> it's likely that implementations of even the most basic functions will behave in subtly different ways across languages

Stares at C/C++


Apparently due to Sam's ego and wanting to use OpenAI to achieve a Musk-like cult status: https://old.reddit.com/r/OpenAI/comments/17xoact/sam_altman_...


Source: trust me bro, I'm a random reddit user.


It's harder to tell now, but from a timeline of events, that supposed OpenAI insider on Reddit had key details of the 'palace coup' that reputable journalists like Kara Swisher (and now a primary source of Greg Brockman) have largely confirmed.

I think the Reddit user is legitimate.


unfortunately, that is also how reddit roleplay works too.


front-running scoops from journalists?


yeah, often times that is the case. Because reddit is antonymous and free, people can and do make all kinds of claims, some of which come true. If you have 500 people making up different claims, it is likely that one of them will come true, even if that person was guessing/lying.

Im not saying that real leaks dont happen, but that pre-running the news is really weak evidence the on reddit, unless the claims are incredibly specific.

I dont think there was anything they said on reddit that was very specific, and several posters in the HN thread speculating the same thing, just without claiming to be "in the room".

It is kind of like publishing multiple weather predictions, and then claiming credit for the one that is correct.


The consistent detail about the store breaking the camel's back sounds rather specific. I guess time will tell whether more claims by this user hold up.


>I guess time will tell whether more claims by this user hold up.

Thats the point, it doesnt matter if they do or dont.

If you read through their post history, they have a comment or sentence for almost every possible reason:

Difference in vision

forcing development to fast

Didn't listen to safety concerns

Lying about finances

Making deals without the board

Hiding financial matters.

It is like someone is rattling off the top 10 guesses for the reason from HN or predication markets.

Now if they said something like Ilya Sutskever called Sam at 1215 on hangouts to tell him, That would be a be pretty solid proof.


The user just made some new comments that lack credibility. I think that seals whatever doubts I had.


yeah, the new comments look like more of the same to me. A lot of not-so-subtle self aggrandizement, vague claims, and of course a "big secret" that they cant share.

They could still be some kind of insider, but they never added any new information beyond what was publicly speculated, and clearly have an axe to grind.

There is something about reddit that just attracts and promotes a lot of juvenile attention seeking behavior. My default assumption is that content is fictional unless it passes a pretty high bar.


No that's horseshit since it does not constitute a valid legal reason for his removal nor it's inline with their blog post. They would get sued out of their a* if they acted based on this.


I mean Ilya said it himself in a roundabout way https://twitter.com/ilyasut/status/1707752576077176907



Came here to say this. As soon as I saw the first example on the linked web page, I immediately thought of C++. Reinventing the wheel.


The Computer Language Benchmarks Game has C++ outperforming Rust by around 10% for most benchmarks. Binary trees is 13% faster in C++, and it's not the best C++ binary tree implementation I've seen. k-nucleotide is 32% faster in C++. Rust wins on a few benchmarks like regex-redux, which is a pointless benchmark as they're both just benchmarking the PCRE2 C library, so it's really a C benchmark.

> because C++ makes it laughably easy to write incorrect code

I was going to ask how much you actually program in C++, but I found a past comment of yours:

> I frankly don't understand C++ well enough to fully judge about all of this


> Rust wins on a few benchmarks like regex-redux, which is a pointless benchmark as they're both just benchmarking the PCRE2 C library, so it's really a C benchmark.

The Rust #1 through #6 entries use the regex crate, which is pure-Rust. Rust #7[rust7] (which is not shown in the main table or in the summary, only in the "unsafe" table[detail]) uses PCRE2, and it is interestingly also faster than the C impl that uses PCRE2[c-regex] as well (by a tiny amount). C++ #6[cpp6], which appears ahead of Rust #6 in the summary table (but isn't shown in the comparison page)[comp], also uses PCRE2 and is closer to Rust #7.

[comp]: https://benchmarksgame-team.pages.debian.net/benchmarksgame/...

[detail]: https://benchmarksgame-team.pages.debian.net/benchmarksgame/...

[rust7]: https://benchmarksgame-team.pages.debian.net/benchmarksgame/...

[c-regex]: https://benchmarksgame-team.pages.debian.net/benchmarksgame/...

[cpp6]: https://benchmarksgame-team.pages.debian.net/benchmarksgame/...


I mean, it's outperforming C as well in that particular benchmark.

Lies, damn lies, and benchmarks?

I can at least say, the performance difference between C, C++, and Rust, is splitting hairs.

If you want to write something performant, low level, with predictable timing, all three will work.

I'm spending a lot of time building projects with Rust & C++ these days. The issue/tradoff isn't performance with C++, but that C++ is better for writing unsafe code than Rust.

https://www.p99conf.io/2022/09/07/uninitialized-memory-unsaf...


Yes, all 48 of them.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: