Hacker News: taway1237

New comment by taway1237 in "ChatGPT with voice is now available to all free users"

taway1237 — Wed, 22 Nov 2023 10:46:35 +0000

I don't know how people use ChatGPT at all. It confidently hallucinated answers to 4 out of 5 my latest "real" questions, with code examples and everything. Fortunately with code I could easily verify the provided solutions are worthless. Granted I was asking questions about my niche that were hard enough I couldn't easily Google them or find a solution myself, but I think that's the bar for being useful. The only thing it got right was finding a marketing slogan.

New comment by taway1237 in "Uber slashes fees in Bangladesh as drivers keep taking rides offline"

taway1237 — Tue, 21 Nov 2023 19:54:46 +0000

>You call a number and a cab turns up and you pay by the meter or an agreed upon price. How did it work then?

Badly. I was scammed multiple times in different countries (I travel a lot), including my own. And there's a lot of uncertainty and stress involved (no guarantee you share a language with your driver, in case they need to ask any questions). And I hate talking over the phone.

As long as that's an option I'm staying with Uber.

New comment by taway1237 in "A secret deal let Spotify bypass Android's app store fees"

taway1237 — Tue, 21 Nov 2023 18:13:58 +0000

On PC you can just mute the music and play anything you want in the background. So there's no need for a special support.

New comment by taway1237 in "Nothing's iMessage app was a security catastrophe, taken down in 24 hours"

taway1237 — Tue, 21 Nov 2023 15:25:57 +0000

Yes, let's give corporations more power. Image bubble colour indicate the protocol you're using, and nothing more. It don't - and can't - guarantee security (many people learned this lesson with https).

But sure, let's let corporations super-copyright the colour green.

New comment by taway1237 in "Sam Altman, Greg Brockman and others to join Microsoft"

taway1237 — Mon, 20 Nov 2023 09:12:50 +0000

>corporate interest is automatically a cartoon supervillain

Not a cartoon villain. A paperclip maximizer.

New comment by taway1237 in "Rational Ritual: Why everyone needs to know what everyone knows (2020)"

taway1237 — Sun, 19 Nov 2023 21:18:59 +0000

I would call most intelligent space aliens from movies "deeply human". Which makes sense, they were created by humans to tell stories to other humans. So I guess being human is not a prerequisite to being "deeply human". I wonder if real space aliens, if they exist, are in fact like this. The closest earthly thing we have to intelligent aliens - octopi - are not.

New comment by taway1237 in "Rational Ritual: Why everyone needs to know what everyone knows (2020)"

taway1237 — Sun, 19 Nov 2023 21:13:25 +0000

I wanted to say that this "obviously" means "5 hours median" i.e. 50%, but I just realised my managers and clients would almost always understand this as "I swear on everything I hold dear this will take 5 hours and not a single minute more" (which is why I avoid sharing my estimates publicly). Funny how it works. Great observation, thanks.

New comment by taway1237 in "OpenAI's board has fired Sam Altman"

taway1237 — Sat, 18 Nov 2023 17:09:03 +0000

I want you to be right. But why do you think you're more qualified to say how to make AI safe than the board of a world-leading AI nonprofit?

New comment by taway1237 in "Privacy is priceless, but Signal is expensive"

taway1237 — Fri, 17 Nov 2023 14:31:52 +0000

So all the programmers who work there should live on thin air? I agree that ideally the management should not be there for profit, but come on, the salaries are not even that crazy. I suspect FAANG key employees in that area easily earn multiples of that.

New comment by taway1237 in "Privacy is priceless, but Signal is expensive"

taway1237 — Fri, 17 Nov 2023 14:26:52 +0000

>I think I personally know quite a few people that could do at least as good a job as Signal has at building a messenger app + platform. No offense, but this isn't exactly rocket science.

They are building a secure communicator that a normal person can reasonably use - and succeeding. Something nobody else before them managed to pull off. If this isn't rocket science I don't know what is. Not to mention that they pioneer cryptographic protocols in this area, which other messengers later use.

>This only makes sense if you ignore the world outside the Bay area and assume it's a talentless wasteland.

I'm also from Europe (and love it, despite its flaws) but this comes off like whining. If it's really so easy, maybe the smart people here should create their own Signal and reap that overinflated salaries, what do you think?

Or maybe smart people are not enough and you also need VCs, reasonable taxes, laws... Oh btw, did you hear about those plans of EU to get rid of E2E encryption?

New comment by taway1237 in "Google News is shutting down purchased magazine content, offering refunds"

taway1237 — Fri, 17 Nov 2023 13:59:53 +0000

1. Piracy is theft.

New comment by taway1237 in "U.S. struck with record 25 separate billion-dollar disasters so far this year"

taway1237 — Mon, 13 Nov 2023 18:13:33 +0000

I have issue with headlines like this. Given inflation it's actually expected that every year is "record", in non-inflation-adjusted dollars. Same goes e.g. for number of millionaires, house prices, etc.

New comment by taway1237 in "Dungeon KeeperFX 1.0"

taway1237 — Sat, 11 Nov 2023 17:04:10 +0000

Answering your question directly: RE is 60% of my daily work, and I don't use and don't know anyone who uses LLMs for RE. One reason is that you lose interactivity if you process code using something other than your decompiler of choice, and it's a huge problem - you want to stick to your RE "IDE". The other is that, at least in my line of work, I don't need perfect code, just to understand what's going on and maybe a few critical functions.

New comment by taway1237 in "Typing fast is about latency, not throughput"

taway1237 — Fri, 10 Nov 2023 14:19:58 +0000

Lucky you. In my previous job I had to work many hours a day over SSH or even RDP on servers an ocean away (servers in the SF region, I worked remotely from europe).

New comment by taway1237 in "Typing fast is about latency, not throughput"

taway1237 — Fri, 10 Nov 2023 13:46:31 +0000

Only when typing english, I assume?

New comment by taway1237 in "i386 Assembly Language trick for storing data in .text"

taway1237 — Thu, 09 Nov 2023 14:45:41 +0000

I disagree. For me it's useful mostly for position independent shellcode prologue, which has no sections to speak of, and may get embedded in a "normal" executable or something that is not an executable at all (useful in a bootloader, or for injecting code to another process, or self-relocating code, etc). I use this "trick" all the time and I never felt the need to mess with a linker for this.

But it's a good hint, I hope I didn't sound overly negative.

New comment by taway1237 in "i386 Assembly Language trick for storing data in .text"

taway1237 — Thu, 09 Nov 2023 09:54:21 +0000

Is this some obscure feature of Windows? In my experience, while code sections are almost never writable, they're always readable.

New comment by taway1237 in "i386 Assembly Language trick for storing data in .text"

taway1237 — Thu, 09 Nov 2023 09:52:18 +0000

The article is about call+pop "trick" in assembly, linker is not relevant here.

New comment by taway1237 in "Euclid's First Images"

taway1237 — Tue, 07 Nov 2023 21:38:54 +0000

>Astronomers demonstrated that galaxy clusters like Perseus can only have formed if dark matter is present in the Universe.

Any MOND people here to comment what they think about Euclid? I always enjoy reading MOND speculation here on HN (even though I don't know enough to have an informed opinion myself).

New comment by taway1237 in "'Ray, this is a religion': How Bridgewater lost two top hires"

taway1237 — Tue, 07 Nov 2023 21:31:21 +0000

>It's always funny when some rag tries to say "X declined to comment". No one is going to comment on some random crap you publish on what's a glorified blog

Strong disagree. I like when people writing articles at least give the other side a chance to comment. It makes me feel the piece is less obviously biased. Of course it can be abused by giving the other side just a dew hours to comment (so basically no real chance), but it's still better than nothing.