AskHN Digest

1. Ask HN: Is Anyone Else Tired of the Self Enforced Limits on AI Tech?

Top comment by Roark66

For anyone who actually used those models for more than few days and learns their strengths and weaknesses it is completely obvious all this talk of "societal impact", or as you called it self imposed limits are 100% bulls**. Everyone in the field knows it.

99% of those using this tactics use it to justify not releasing their models to avoid giving competition a leg up(Google, openAI) and to pretend they are for "open research". As I said this is 100% bull.

The remaining 1% are either doing this to inflate their egos ("hey look how considerate and enlightened we are in everything we do!"), or they pander to media/silly politicians/various clueless commentators whose level of knowledge about this technology is null. They regurgitate the same set of "what ifs and horror stories" to scare the public into standing by when they attempt to over regulate another field so they can be kingmakers within it(if you want an example how it works look at the energy sector).

All this sillyness accomplishes is to raise a barrier to entry for potential commercial competition. Bad actors will have enough money/lack scruples to train their own models or to steal your best ones regardless how "impact conscious" your company it.

Now, I don't claim everyone should be forced to publish their AI models. No, if you spent lots of money on training your model it is yours. But you can't lock all your work behind closed doors and call yourself open. It doesn't work like this. One important point is that there is value even in just publishing a paper demonstrating some achievements of a proprietary model, but if the experiment can't be reproduces based on description given that is not science and for sure it is not open.

2. Ask HN: Advice that changed your life?

Top comment by javajosh

My mother told me to not care what other people thought. She also told me that words can't really hurt people ("sticks and stones..."). She also believed its "mean" to have any preference about people, calling it a version of "keeping score". In fact, "keeping score" was one of the worst things you could be accused of in my household.

Her advice fails to accept the reality of social and emotional needs. It also fails to address the (sometimes severe) downside of non-conformance. This advice made me an Outsider whether or not I wanted to be. I now believe that these were tools she formed to get her through her own family and social trauma, which she falsely assumed would be generally useful, and did her best to arm her children with her best tools. I also believe there are upsides to the approach, but the trade-offs are real and cannot and shouldn't be swept under the rug. I imagine others had similar experiences with religious parents.

Advice is dangerous. It has built in moral hazard. Advice is too often given in well-meaning ignorance and pride - but ultimately advice gets someone else to test your hypothesis for you. So if you've not gotten any life-changing advice, be glad!

3. Ask HN: What are some of the best books you have read in 2022?

Top comment by plants

I’m sure not everyone came here to hear tech book recommendations, but I will add another vote for Designing Data-Intensive Applications by Martin Kleppman. It’s one of the best tech reference books I’ve ever read. It manages to explain SO much while requiring so little prior knowledge from readers.

Another book that is relatively new that I loved was Designing Machine Learning Systems by Chip Huyen. I worked in productionizing ML systems for 3 years and this book equips you with exactly what you need to do so. It does a great job of explaining the whole ML modeling pipeline and some of the commonly overlooked nuances that can cause your models to fail spectacularly in production. I will be referencing this book for years to come.

4. Ask HN: Do people not have hobbies anymore?

Top comment by thiago_fm

Hey CM30, I feel you are using too much internet.

I started writing tweets with Ruby tips and engaging with people for sometime and I've noticed most people tend to find a way to monetize their activity. I only do it for the pleasure of learning with the community. I don't care about writing ebooks or making people pay for things they don't need.

But I can definitely conclude that on the internet, most people are looking to make big bucks. But have you tried going climbing or doing outdoor activities? There is always plenty of amateurs that never dream of becoming a professional climber or anything like that.

Also depends on where you live, but I live in Europe and here you see people having hobbies and doing random stuff everywhere you go, people just drinking a coffee and staring at the sky.

It could be that this is what you need. Disconnect.

5. Ask HN: What is your fav “I can't believe in 2022 this still doesn't work right”

Top comment by wizzwizz4

Latency.

Back in the day, you flicked your computer's on switch, and it turned on. (And then it took ten minutes or so to load software from the tape drive – much faster from floppy, but still far from instant.) Now, our processors and storage devices are millions of times faster, and computers can take minutes just to start. Software needs to load itself into memory, then initialise the initialisation routines, then actually initialise, then you have to wait for the data to populate, then it's still slow until the cache has warmed up.

It's 2022 and if I want to play a rhythm game, I have to calibrate my input device because, since the days of PS/2, keyboard latency has gone up so much that I can hear the delay between the button click and the computer's reaction.

It's 2022, and when I draw something on one device, it takes several seconds for the drawing to make its way to the other, even if they're in the same room, because the data's being sent through four layers of format conversion in JavaScript, and the traffic's being sent half-way 'round the world.

It's 2022, and "blazing fast" means "slower, and probably more buggy, than Windows 95".

6. Ask HN: Are there recorded instances of people being framed through hacking?

Top comment by jimlongton

Yes, look up the Bhima Koregaon case in India. Indian police used Israeli spyware to hack the phones of lawyers, human rights activists and critics of Modi. They also used phishing and other malware to plant terrorist material and then imprisoned them.

> In Wilson’s case, a piece of malware known as NetWire had added 32 files to a folder of the computer’s hard drive, including a letter in which Wilson appeared to be conspiring with a banned Maoist group to assassinate Indian prime minister Narendra Modi.

[1] https://www.washingtonpost.com/world/2021/07/20/indian-activ...

[2] https://www.wired.com/story/modified-elephant-planted-eviden...

7. Ask HN: Stripe holding funds for 120 days for no reason

Top comment by johnhaddock

I’m the person leading the project to make sure that people have fewer bad experiences with Stripe. I’m not sure what is driving the uptick in posts on the subject to HN in particular (obviously we pay attention to broader online discussion in addition to monitoring our support systems, though we know it creates an incentive for people to publicize their situation).

On the topic of Stripe and these kinds of incidents more broadly, there’s a lot to say, but here are a few pieces of context that are probably relevant:

- We are a giant distributed bounty system for people to find interesting and scalable ways to defraud us.

- We’ve seen significant upticks in certain kinds of fraud over the past couple of months. When businesses default, Stripe takes on the loss. It’s worth noting that certain kinds of fraud, like card testing, can also have significant collateral costs for legitimate Stripe businesses, and our systems and processes are not only to protect Stripe itself.

- We are far from oblivious to the harm that mistakes in our systems can cause. (I interact with a lot of these cases personally.) One of my highest priorities is creating better appeals flows for when we’re wrong.

- We’ve shipped 7 substantial improvements just in the last 10 days that should meaningfully reduce the occurrence of false positives.

- Publicly-described facts of specific cases don’t always match the actual facts. Stripe is sometimes just wrong. (We made some mistakes that I feel bad about in one recent case and we ended up bringing the company’s founders to an all hands last week to make sure we learned as much as possible.) But users do also sometimes publicly misrepresent what’s going on. We’re also restricted by privacy rules to not share specifics in those cases.

- Stripe works with millions of businesses and we see all kinds of “rare” failure modes fairly frequently. (Disputes between staff at a business, business impersonation, businesses that start legitimate and go bad, and so on.)

- I’m working on a post to share some of our broader philosophy + policy changes that I hope to publish before the end of this year. In that, I’m also hoping we can share some relevant metrics. If HNers have any suggestions for things that might be useful to see covered (though obviously certain things can’t be publicly disclosed), feel free to suggest them.

Ultimately, we work hard to be worthy of the trust of businesses across the internet, and my personal mandate (supported by many others, from our cofounders down) is to find effective new ways of making mistakes less likely. “Uniformly good support at scale, in a highly adversarial environment, with very financially-motivated actors” is not easy, but I’m pretty confident that we can make a lot of progress.

It goes without saying we're working on a review of OP situation. I’m happy to take general questions as well. You can also always reach me directly at jhaddock@stripe.com.

8. Ask HN: I've Built a DHT Torrent Sniffer and Search Engine. Should I Release?

Top comment by boramalper

I had been working on this successfully for a couple years in the past before I got tired of it and moved on. I still think it's a magnificent idea, to be able to host your own torrent site and to decentralise the last centralised bit of BitTorrent.

https://github.com/boramalper/magnetico

9. Ask HN: How to overcome job search exhaustion?

Top comment by wincy

I don’t know OP, I’m a tech lead who has done his fair share of interviews and based on your few comments on HN there seem to be some red flags with how you’re viewing the interview process. Saying “one person derails things because I’d outshine them” seems to externalize the blame. I don’t know that I’ve ever worked with any engineer who thinks they’d be worse off by hiring an extremely intelligent engineer. I’d love to be “outshined” as it’d make my life easier to have someone competent writing code with me and teaching me new things. I have a strong suspicion “I’m too smart” isn’t the issue here.

Also, do you honestly think this has been the case at over forty places? That in dozens of separate cases you weren’t the problem, the interviewers were? This seems like an extraordinarily hostile and pessimistic worldview and I’d guess it’d come across in an interview. Eventually you need to do some soul searching into what’s going awry in the interviews. I have a good friend who is a software engineer and is autistic. He speaks about his interviews similarly to what you’ve said. Maybe ask a friend about what they might think is the problem. Be prepared though, because often if you solicit honest feedback from friends you won’t like what they have to say.

Also why are you only applying at places that do leetcode style interviews? There’s plenty of places where that’s not the case, especially if you work outside of tech companies. There’s plenty of great Fortune 500 companies that will expect a 9-5 and are doing good engineering out there. Most of them are hiring remote. Although if you’re expecting FAANG salaries it might be harder to find that type of place.

I wish you the best and hope you get to a better place. Good luck!

10. Ask HN: Recommend me a Blender course

Top comment by alomaki

Blender Guru's Donut tutorial is regarded as the "Hello World" of Blender. Long tutorials will start getting boring after a certain period of time. So I would suggest following a couple of random Ducky 3D's abstract designs after this one. After that follow the OG IanHubert's lazy 1 minute tutorial.It contain only the core concepts remaining you have to explore and find out. Checkout Polyfjord also.