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

Ask HN: What interesting problems are you working on?

news.ycombinator.com

Post reply on HN
Print view View original

171–180 of 633 posts

Re: Ask HN: What interesting problems are you working on?

#171

Post by smcn »

doix wrote: post #161
smcn wrote: post #130

Earlier quoted context omitted.

Good question! It's a manual process of discovery. As in, I actually go out and hunt down groups talking about stocks and if it looks interesting, I'll add it into the algorithm. I'll then analyse the data over a period of a week or two and determine whether it's worth continuing. If it is, I move it to production and it'll perform as the rest do. If a community is not publicly visible (discord, telegram, etc) I'll a…

Wait, I thought the whole point was to find pump and dumps early. I'm not sure what you're trying to do now. The most interesting (for me anyway) would be to figure out what is going to be on Reddit before it appears by analyzing spam/pump and dump groups etc. If you're analyzing Reddit after the fact, isn't that kinda late?

No, not pump and dumps. Just stocks that a large amount of people are about to buy into. A pump and dump is a coordinated effort, whereas I'm looking for something more organic.

It might be late in that you might make less money than you would if you were in those groups but I can't say that I'm unhappy with what the algorithm is currently finding. It's primarily built for day trading, so you're in and out of trades that same day. There are times you're out within 30 minutes. The goal is near daily compounding, which leads to higher gains than holding for X number of days.

It has been recording data since 9th of August 2021 and is averaging 4.5648% across 340 stocks. Since Jan 1st this year, it's averaging 5.1779%, and I would like to think that I can get that number higher.

Note that these are perfect values, using the open price and the highest price of the day, they're unachievable consistently, I use them to understand how much potential for profit we could've had.

Re: Ask HN: What interesting problems are you working on?

#172

Post by atlasunshrugged »

justicz wrote: post #128

Robots that build large-scale solar farms! My friend and I started working on this last year and ended up starting a company. We moved to the bay area and did YC last summer. We've been spending a lot of time on solar construction sites and did our first demo (part of the material handling system) on a site in March. Now we're working on a robotic factory + our team is up to 5 folks.

Awesome, think I saw your launch post and sounds like an awesome project. How has the experience been with permitting and regulatory issues (not technical)? Is there a valid strategy in building at the border of hard to build states and selling electricity back to them (e.g. building a massive facility in Nevada and selling the energy to CA when they inevitably suffer another crisis to the grid)?

Re: Ask HN: What interesting problems are you working on?

#173

Post by scubakid »

I'm building a privacy-friendly personal finance simulator that doesn't ask to link your accounts (projectionlab.com), and can run Monte Carlo simulations in your browser and keeps data client-side unless you choose otherwise. I'm hoping to spin up a self-hosted version soon as well.

The overall design ethos/constraints have posed some interesting challenges along the way, but so far I've had good luck developing with the following: Vue.js, Vuetify, vuex, Chart.js, Threads.js for orchestrating web workers, vue-router, a bunch of smaller libraries, Paddle, Firebase, and some Google Cloud Functions (the latter only come into play for those that upgrade + choose to enable cloud sync).

For now, it's still a side project :)

Re: Ask HN: What interesting problems are you working on?

#174

Post by atlasunshrugged »

throwawayacc2 wrote: post #25
Test0129 wrote: post #12

Earlier quoted context omitted.

> Why artistic productions? Well, because it’s the most liberal business code for expenses. > As a software company, I can expense travel for some things, the most useful for me being conferences. But that means I need to buy the conference ticket even if I’m not interested and can only expense accommodations for the duration of the conference. That’s good but a bit restrictive. > Now, as a artistic production compan…

It doesn’t. I’m in the EU. There seems to be a false idea that only the USA has tax loop holes. I assure you. It’s not the case. And tax heavens are also way easier to use as a EU citizen because there’s no global income declaration required.

Did you ever explore Estonian e-Residency to open a biz and run it there? Curious your thoughts/experience about that.

Re: Ask HN: What interesting problems are you working on?

#175

Post by memset »

pwpw wrote: post #170
memset wrote: post #133

Revenue recognition accounting. (Is that not interesting to you? It's super interesting to me!) There's this whole world of rules that accountants follow to properly record revenue. These rules are different for physical products, subscriptions, contracts, and even rules for what to do when you bundle these things together. And yet, the only two options for companies to crunch all this data is to use an Oracle produc…

Hey I’m a CPA (former big 4 accountant) that’s taught myself how to code. Interested in learning more if you want to share: contact+hn@winstoncooke.com

I’ll reach out! Also looking at your GH, you might be interested in this blog post I wrote https://blog.journalize.io/posts/an-elegant-db-schema-for-do...

Re: Ask HN: What interesting problems are you working on?

#176

Post by elteto »

scubakid wrote: post #173

I'm building a privacy-friendly personal finance simulator that doesn't ask to link your accounts (projectionlab.com), and can run Monte Carlo simulations in your browser and keeps data client-side unless you choose otherwise. I'm hoping to spin up a self-hosted version soon as well. The overall design ethos/constraints have posed some interesting challenges along the way, but so far I've had good luck developing wit…

This sounded interesting and I want to follow up on it in a while to see what progress you've made. However, I know I will forget in the future to actually follow up so I wanted to sign up for email updates or even a newsletter if you had one, but I couldn't find one from your landing page.

Re: Ask HN: What interesting problems are you working on?

#177

Post by PainfullyNormal »

fikama wrote: post #105
PainfullyNormal wrote: post #79

I'm going back to basics and learning to write a compiler in the dumbest way possible. I'm starting with a basic assignment (a = 3) and reverse engineering what it would take to turn that into a working executable. Then I'll add more features, lather, rinse and repeat. I'm hoping to eventually turn it into a series of blog posts so everyone can see just how little I remember from college.

I like it so much, reading your comment make exited to reading your blogpost. In moments like this I wish there was sth in hacker news that would allow tracking progress of e.g. your project. In twitter you can just follow people, reddit has remindme thing.

If you're interested, I created a substack I'll be publishing the first few posts to. You should be able to subscribe there.

https://painfullynormal.substack.com/

Re: Ask HN: What interesting problems are you working on?

#178

Post by scubakid »

elteto wrote: post #176
scubakid wrote: post #173

I'm building a privacy-friendly personal finance simulator that doesn't ask to link your accounts (projectionlab.com), and can run Monte Carlo simulations in your browser and keeps data client-side unless you choose otherwise. I'm hoping to spin up a self-hosted version soon as well. The overall design ethos/constraints have posed some interesting challenges along the way, but so far I've had good luck developing wit…

This sounded interesting and I want to follow up on it in a while to see what progress you've made. However, I know I will forget in the future to actually follow up so I wanted to sign up for email updates or even a newsletter if you had one, but I couldn't find one from your landing page.

Sounds like I need to make the newsletter more discoverable. Currently it only pops up if you go to the blog or if you do the onboarding wizard or sandbox. I'll take a note to add it to the homepage as well.

Re: Ask HN: What interesting problems are you working on?

#179

Post by bqmjjx0kac »

thfuran wrote: post #120
PontifexMinimus wrote: post #73

Earlier quoted context omitted.

How will you prevent people from using this technology to create DIY pathogens? If the answer is "I won't" or "it'll be open source, I won't be able to", consider doing something less harmful instead.

How far down this road do you go? Someone engaged in bioweapons manufacturing probably uses laptops and chairs too.

The difference is that those already exist. I think concern about the development of new technology is valid.

Suppose nuclear weapons did not exist and someone was concerned about the Manhattan project — would you say that chairs enable people to build regular bombs, so we shouldn't worry about nuclear bombs?

Re: Ask HN: What interesting problems are you working on?

#180

Post by hazrmard »

I am working on fusing classical control with reinforcement-learning based control in drones.

The good thing about RL is that it is adaptive and proactive - it tunes control from data as it streams and it learns to act with delayed feedback in mind.

The bad thing is that it can be sample inefficient. It needs lots and lots of data. That is not good in a time-critical case such as when faults occur. I've come to enjoy pushing back on completely black-box control of systems, foregoing the advances of decades of theory. That helps me scale back complexity of the system if only a little domain knowledge is sprinkled on the solution.

So, first, I started using transfer learning to re-use existing policies from similar situations. However "similar" is a very open problem in the domain. I then thought that we can exploit insights we get from classical control of linear systems into these more complex domains. If, for example, we can make some simplifying assumptions (locally linear, there's some global optimum etc.). So far I am getting promising results. But there is always some data wrangling involved. I am hoping to use that in my dissertation later this year. (If anyone's hiring, feel free to reach out).

Post reply on HN