Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ask HN: What's a promising area to work on?

news.ycombinator.com

Post reply on HN
Print view View original

851–860 of 947 posts

Re: Ask HN: What's a promising area to work on?

#851

Post by al_chemist »

echelon wrote: post #675
ncmncm wrote: post #542

Earlier quoted context omitted.

The most pressing problem for humanity is global climate disruption. Arguably, working on anything that doesn't help there is not just wasted effort, but actively harmful. Wind, solar power, energy storage still have huge room for improvement. Direct solar or wind -> liquid fuel will be essential to adapting fast. Windmills can drive production of liquid ammonia on farms (needing only air and water inputs), for use d…

> The most pressing problem for humanity is global climate disruption. Our climate and ecology are certainly at risk, but I think the biggest threat in the world right now is the rapid rise of fascist China. They're challenging the notion that free speech and democracy are required for capitalism and economic gain. China has grown so emboldened under President Xi that they're no longer content to just alter or buy ou…

>> Without massive progress in the next decade, civilization will probably collapse by 2035.

> That's a little bit overblown. Want to make a bet on longbets.org? I'll happily donate to a green cause.

"If civilization collapse by 2035, I vow to donate 3 flints and 10 sticks to a green cause."

Re: Ask HN: What's a promising area to work on?

#852

Post by al_chemist »

secabeen wrote: post #508
jakemal wrote: post #404

Earlier quoted context omitted.

It reminds of the article that was posted yesterday ( https://news.ycombinator.com/item?id=21315942 ) that said the biggest disruptors are the technologies that start out looking like toys. An HN commentor mentioned (which I now can't find) that what differentiates actual toys from big technologies that start out looking like toys is the fact that the fad never passes despite looking like a toy. Crypto has seemed lik…

If someone can figure out how to do crypto without burning through thousands of CPU-hours of compute, they would have a real winner on their hands. Of course, execution is also a key factor in success. There are lots of good ideas that were had by multiple people, but only one successfully executed.

> do crypto without burning through thousands of CPU-hours of compute

Imagine how video games could grow if only somebody invented how to make games that doesn't burn through thousands of GPU-hours of compute.

Re: Ask HN: What's a promising area to work on?

#853

Post by al_chemist »

screye wrote: post #273

ML for healthcare. The whole industry is so far behind everything else, it is not even funny. Problem is, Just like Airbnb or Uber, the biggest road blocks are not technological, they are organizational and political. It is entirely possible, we might need some major privacy / security related leaps in ML to convince the medical industry to adopt it fully. ML in movies and rendering. Deep Fakes is just an amateur's t…

> give a company rights to use a 3d render of you in it

There's reason why actors earn money and it's not due to perfect face or body.

Re: Ask HN: What's a promising area to work on?

#854

Post by ncmncm »

1337biz wrote: post #827
SeaSeaRider wrote: post #727

Earlier quoted context omitted.

“Without massive progress in the next decade, civilization will probably collapse by 2035.” This is deathcult type rhetoric which has no scientific basis whatsoever.

I just wish there would be a way to hold these doomsday preachers responsible. In the end they will all shrugg it of with that smirk of 'oh back then we didn't know better'.

If you know how to manage involuntary migration of 100M-2B population from newly uninhabitable land to established sovereign territory without triggering global thermonuclear warfare, do speak up.

Re: Ask HN: What's a promising area to work on?

#855

Post by nnd »

krageon wrote: post #815
teknologist wrote: post #786

Earlier quoted context omitted.

Two words: Signal Protocol

It's an american development by a company that refuses to release actually open software. Perhaps it's good in theory! As far as I am aware there is no open implementation of it and it is also inside a problematic regime and therefore just as suspect.

The clients are available under GPLv3.

Re: Ask HN: What's a promising area to work on?

#856

Post by mellosouls »

K0SM0S wrote: post #709
mellosouls wrote: post #697

Earlier quoted context omitted.

The idea of cycles in social history is as old as the hills. It's possible your claim that he drew from Strauss-Howe is correct, but while they may have packaged the idea in a distinctive way, it has a long history. https://en.m.wikipedia.org/wiki/Social_cycle_theory

Oh indeed! Thanks for the precision, well worth noting. The particular arrangement of The Fourth Turning is adamantly specific though (and quite famous in sociological circles, iirc it's where our modern concepts of "millenials" or "gen X" etc. come from); it's hard to mistake it for anything else. I don't doubt for one second that Greene knows it, or that he'd expect a sophisticated audience (such as Googlers) to kn…

Thanks I didn't realise the naming originated with them. wrt pseudo science the whole social cycles subject is interesting and sometimes plausible enough to read about even if you take it with a large pinch of salt.

Re: Ask HN: What's a promising area to work on?

#857

Post by eitland »

ryanlol wrote: post #777
eitland wrote: post #769

Earlier quoted context omitted.

1. They are vey different. Those who arent aware of the large differences shouldn't comment as if they are they are subject experts. Telegram might be full of holes but it is a very different thing to Slack. (I use both for different things.) 2. I have my own issues with Telegram (the refusal to introduce sustainable funding, pushing of crypto currencies, possibly also marketing it as more than it is) but except that…

>1. They are vey different. Those who arent aware of the large differences shouldn't comment as if they are they are subject experts. Telegram might be full of holes but it is a very different thing to Slack. (I use both for different things.) This is just bullshit. Slack does TLS, Telegram does their own weird TLS replacement. In both cases essentially zero conversations are E2EE. As far as privacy and security goes…

This is getting tiresome and I don't think I'll move ryanlol, but for everyone else:

One big difference between slack and Telegram is that Slack can and will give my data to the administrator at work if they ask for it.

ryanlol seems to be thinking only about the last hop from their data center to the device. In practice security is about a whole lot more than just checking of the coolest boxes: WhatsApp for example has had e2e encryption for a while now all while uploading every last message of every conversation to datacenters that are easily accessible for anyone with an American subpoena.

That is totally OK, I'm posting this from my old account and I don't have anything to hide from American authorities but hardly compatible with what certain people here think that just because something is e2e encrypted then it safe and good.

So why is Telegram bulletproof. That was a trick question : it isn't. But we should stick to the facts instead of trying to tear down a strawman.

Re the rest of ryanlols post: as I've mentioned before most people here don't understand Russian, yet ryanlol keep posting old Russian posts. I don't know why.

Re: Ask HN: What's a promising area to work on?

#858

Post by companyhen »

lazzlazzlazz wrote: post #276

Crypto. And no, I don't mean applying blockchain to everything, or internet money that just goes "up and to the right". I mean the new applications of distributed systems research and cryptographic primitives that allow for highly composable, highly trustworthy, permissionless, autonomous machines. MakerDAO, Arweave, the whole field of open finance (aka "DeFi"), and so many others collectively are very likely to chan…

Happy to see Arweave mentioned here, their dev community is super strong and they're doing a lot right from what I see.

Re: Ask HN: What's a promising area to work on?

#859

Post by eitland »

ryanlol wrote: post #816
krageon wrote: post #815

Earlier quoted context omitted.

It's an american development by a company that refuses to release actually open software. Perhaps it's good in theory! As far as I am aware there is no open implementation of it and it is also inside a problematic regime and therefore just as suspect.

Yeah, the Kremlin backed startup that was already caught shipping blatant backdoors is so much better https://habr.com/en/post/206900/

1. The man behind Telegram cannot even safely visit Russia after people with friends in high places stole his previous startup there or so the story goes.

2. As have been pointed out before, while this is really bad if true, but you are pointing to something that happened extremely early in the development of Telegram and later fixed (and as I've mentioned before you keep throwing a link to an old Russian post in a forun where most people have never read a word Russian. I'll add this time that it is almost as you don't want anyone to read it.)

Re: Ask HN: What's a promising area to work on?

#860

Post by eitland »

ryanlol wrote: post #778
johnmorrison wrote: post #775

Earlier quoted context omitted.

E2E encryption is kind of a fucky thing, some suggest it is fundamentally impossible for central web services and even mobile / desktop apps. Basically, if there is a third party involved in the code besides Alice and Bob, the two can never guarantee E2E encryption. Web part is simple, as there are countless ways you can get malicious code delivered from what you think is the correct, safe, web server. Same concept a…

>E2E encryption is kind of a fucky thing, some suggest it is fundamentally impossible for central web services and even mobile / desktop apps "some suggest" How about you actually name some competent people who suggest this? >Basically, if there is a third party involved in the code besides Alice and Bob, the two can never guarantee E2E encryption. Oh, this is a downright insane, dishonest argument. Perfection is imp…

An upvote for that just to prove that we can agree as long as we stick to the facts however harsh they are.
Post reply on HN