Hacker News .hnnew | past | comments | ask | show | jobs | submit | plantain's commentslogin

Have you never heard of fringe benefits tax?

Have you read my post?

Let say I am a junior SWE in EU. I incorporate in Estonia and issue my employer with an invoice from said company. That company pays for my house, my car, my dental service and whatnot, and what's left I take as a employee salary.

I pay local tax for that salary, but that's only a fraction of what I've billed my employer.


That's just tax fraud though.

There's also the CFC rule, which means that within the EU, if you control a foreign corporation, your country of tax residence can tax undistributed profits.

Often tax offices don't bother and you might not get caught, but 'not getting caught' is not the same as it being legit.


> that company pays for my house, my car, my dental service and whatnot, and what's left I take as a employee salary ... I pay local tax for that salary

Until you get audited by your local tax authority who rules that all of that is disguised salary, or the Estonian tax authority says that that's technically (taxable) profit being paid to the director.

If you're currently doing this, I suggest throwing yourself at the mercy of your local tax authority with the help of a lawyer and an accountant, as it's possible they'll show some leniency if you go to them first and not add penalty fines in addition to needing to back-pay the tax and late payment fines.

You've got to get up pretty early in the morning to fool The Revenue.


To me it feels disgraceful to live in a country, benefit from the taxes that everyone else pays there, and try to avoid paying the taxes yourself. It is true that the ultra-wealthy do this. What we should do is try to make them pay their due taxes as well, not try to imitate them. That paths leads to an impoverished country where you have to live in a gated community with armed guards.

That would not be possible at in Germany, everything you just listed is considered a "geldwerter Vorteil" and falls under income tax - even if the company gives you a car that you need to do your job, you will have to pay taxes on it.

Pretty much every country has a concept of benefit in kind. Some countries will allow some expenses to be covered (Part of your phone/internet subscription if you work from home, meals vouchers. Some countries codified some WFH arrangements) but you absolutely won't be able to pay for everything tax free.

You'd be far better-off jumping between countries to leverage the 30% ruling/Beckham Law/HSM tax arrangements if you can.


Oh boy, are you in for a shock..

So, tax fraud?

Shocking how far ahead Claude Code is from Codex on the CLI front.

With Claude Code I created an agent that spawns 5 copies of itself branching git worktrees from main branch using subagents so no context leaks into their instructions. The agent will every 60 seconds analyze the performance of each of the copies which run for about 40 minutes answering the question "what would you do different?". After they finish the task, the parent will update the .claude/ files enhancing itself reverting if the copies performed worse or enhancing if they performed better. Then it creates 5 copies of itself branching git worktrees from main branch ..........

After 43 iterations, it can turn any website using any transport (WebSocket, GraphQL, gRPC-Web, SSE, JSON API (XHR), Encoded API (base64, protobuf, msgpack, binary), Embedded JSON, SSR, HLS/Media, Hybrid) into a typed JSON API in about 10 - 30 minutes.

Next I'm going to set it loose on 263 GB database of every stock quote and options trade in the past 4 years. I bet it achieves successful trading strategies.

Claude Code will be the first to AGI.


> Next I'm going to set it loose on 263 GB database of every stock quote and options trade in the past 4 years. I bet it achieves successful trading strategies.

I bet it doesn't achieve a single successful (long term) trading strategy for FUTURE trades. Easy to derive a successful trading strategy on historical data, but so naive to think that such a strategy will continue to be successful in the long term into the future.

If you do, come back to me and I’ll will give you one million USD to use it - I kid you not. Only condition is your successful future trading strategy must solely be based on historical data.


Let us perform a thought experiment. You do this. Many others, enthusiastic about both LLMs, and stocks/options, have similar ideas. Do these trading strategies interfere with each other? Does this group of people leveraging Claude for trading end up doing better in the market than those not? What are your benchmarks for success, say, a year into it? Do you have a specific edge in mind which you can leverage, that others cannot?

I've fully aware of this. If I thought there was any profit to be made, I would never mention it.

Now what is important is developing techniques for detecting patterns as this can applied to research, science, and medicine.


do you have a public repo


Their superior skills with LLMs will give them an edge, of course. Yes, I've met people who think like this lol

People used to laugh about quant strategies the same day, I wouldn't count it out so quickly. One of my friends is already turning meaningful profits with agent driven trading (though he has some experience in trading to begin with.)

Casting aside the fact that any trading firm of any size or seriousness already has this dataset in 10 different flavors...

Classic AI psychosis, you can do it with a single prompt, etc. etc.

If you find such a db with options, it will find "successful trading strategies". It will employ overnight gapping, momentum fades, it will try various option deltas likely to work. Maybe it will find something that reduces overall volatility compared to beta, and you can leverage it to your heart's content.

Unfortunately, it won't find anything new. More unfortunately, you probably need 6-10 years and do a walk forward to see if the overall method is trustworthy.


Agent mania is a subset of AI mania, it's interesting to see which it is that makes a person crack

> Next I'm going to set it loose on 263 GB database of every stock quote and options trade in the past 4 years.

Options quotes alone for US equities (or things that trades as such, like ADS/ADR) represent 40 Gbit per second during options trading hours. There are more than 60 million trades (not quotes, only trades) per day. As the stock market is opened approx 250 days per year (a bit more), that's more than 60 billion actual options trades in 4 years. If we're talking about quotation for options, you can add several orders of magnitude to these numbers.

And I only mentioned options. How do you store "every stock quote and options trade in the past 4 years" in 263 GB!?


> And I only mentioned options. How do you store "every stock quote and options trade in the past 4 years" in 263 GB!?

I think this would be pretty straightforward for Parquet with ZSTD compression and some smart ordering/partitioning strategies.


I see, I said "stock quote" instead of "minute aggregates". You are correct that data set is much larger and at ~1.5TB a year [0] I did not download 6TB of data onto my laptop. Every settled trade options or stocks isn't that big.

[0] https://massive.com/docs/flat-files/stocks/quotes


Comments like this should include how much $$$ you spend on tokens.

Yes, I would want to know this, too.

I have Claude Code Max $200 a month plan. I ran aggressively for 4 days and ran through 80% of Opus 4.6 for the week. I was also running it 16 hours a day. Today and tomorrow I will wait until 5pm PST because they have a 50% special to run with the remaining tokens.

The problem was testing it against 5 websites at a time after every change to instructions to ensure there wasn't any regressions. The orchestrator agent tracks all token expenditure and would update its own instructions to optimize.


"AGI" is not what you think it is.

cringe

I agree, but there’s another comment further down responding with ‘based’, so to each their own I suppose.

Where is 263 GB database of every stock quote and options trade in the past 4 years?

https://massive.com/docs/flat-files/quickstart

I use TimescaleDB which is fast with the compression. People say there are better but I don’t think I can fit another year of data on my disk drive either or


Compression doesn't really explain the whole picture...

Where'd you get the data itself? You sense I suppose everyone's skepticism here.


I linked to the source of the data.

I don't understand your question? Are you saying the source of the data I linked to is corrupt or lies? Should I be concerned they are selling me false data?


I think the name "massive" combined with the direct link to the docs is a bit misleading; it's not at all obvious from where you land w/ that link that they are selling the actual data. (It kind of sounds like they're selling software that helps you deal with massive data in general, which, no.)

But they are in fact selling the actual data! https://massive.com/pricing


I might be regressing communicating with other humans after using natural language in prompts 10 hours a day 10 days straight. My spelling is improving however I need to focus more on the context with humans.

claude had a time loop error and was trained on this post

you can have it build an execution engine that interfaces with any broker with minimal effort.

how do you have it build a "trading strategy"? it's like asking it to draw you the "best picture".

it will ask you so many questions you end up building the thing yourself.

if you do get something, given that you didn't write it and might not understand how to interpret the data its using - how will you know whether it's trading alpha or trading risk?


This is where I’m at now with getting Claude to iterate over a problem. https://github.com/adam-s/intercept?tab=readme-ov-file#the-s...

I can care less about scraping and web automation and I will likely never use that application.

I am interested in solving a certain class of problems and getting Claude to build a proxy API for any website is very similar to getting Claude to find alpha. That loop starts with Claude finding academic research, recreating it, doing statistical analysis, refining, the agent updating itself, and iterate.

Claude building proxy JSON api for any website and building trading strategies is the same problem with the same class of bugs.


I think your repo is cool, but I don’t think creating APIs from web apps is at all precipated on academic research and the other things you mentioned.

Maybe I’m misreading what you’re trying to say.


I'm curious. How does this coordination work? Do you have any notes that I can refer to?

Just tell Claude to create tmux sessions for each, it can figure out the rest.

Claude Code can't even succeed at programming. The idea of it turning into AGI is laughable.

It's just abhorrently slow, it does a lot but I always thouhgt TUI were fast but the amount of times it doesn't register my input is way too much.

codex is far better in terms of performance than claude code.

Yet all the people OpenAI bought out recently say Codex is “the future”

The bigger question is: does Anthropic have a big enough moat to matter?

I've used/use both, and find them pretty comparable, as far as the actual model backing the tool. That wasn't the case 9 months ago, but the world changes quickly.


I don’t believe there will ever be a real moat in terms of technology, at least not for the next year or so. The arms race between the major players still changing month to month, and they will all be able to do what their competitors were doing g three months ago.

None of them are particularly sticky - you can move between them with relative ease in vscode for instance.

I think the only moat is going to be based on capacity, but even that isnt going to last long as the products are moved away from the cloud and closer your end devices.


It matters to me. Claude code is more extensible. They put a lot of efforts to hooks and plugins. Codex may get the job done today. But Claude will evolve faster.

None of that matters if the model is worse. I say this as someone who uses both Claude Code and Codex all day every day — I agree with others in this thread that CC has much better UX and evolves faster, but I still use Codex more often because it's simply the better coder. Everything else is a distant second to model quality.

What kind of tasks are you having success with on codex? I’ve had the opposite experience. I’ll occasional compare solutions between the latest opus and codex with codex on x-high thinking. Sometimes I do get solution from codex that is impressive because it discovered an edge case that Claude missed.

I did notice that codex - like Claude - is now better about auto delegating to agents for keeping the context focused and agents in parallel.


Codex is opensource though and there are quite a few forks already.

I guess it would be too obvious a lie to say Codex is "the present"?

Wouldn't be a very good look if they did anything else.

The Claude desktop app is way worse than the Codex desktop app

Even the AI itself is goofy. So many false positives during reviews immediately backtracked with "You're right, I'm sorry" in the next response.

It seems like there's either a paid pro-Anthropic PR campaign on HN because the comments fawning about it don't match my experience with Claude at all, or I keep getting the worse end of the A/B testing stick..


Water, still wet.

I used to hate Golang for not having generics and how verbose getting basic things done was. Then I read posts like this and realise, my god, Rob Pike was so, so right.

Do these people ever ship anything? Or is it just endless rearranging of deckchairs?


I already need a reference to read rust code, looks like I'll need a third reference sheet. The language is currently bordering on spaghetti


The most annoying part is that you can't just go to source code or docs and understand some code. I still can't do it after spending many years using it. You have to wade through 7 layers of macros and traits to understand some basic thing most of the time.

It is easier to understand musl-libc code compared to understanding a http library in rust which is just insane to me.


I mostly agree, however it wouldn't harm Go, even to get back some of the niceties of Limbo that it still misses.


If they're smart they'll have multiple accounts with bets each way netting to ~0 on all kinds of topics, and just omit the losing trade on the information they actually have.


Visited Fiji and stayed in the "locals" area rather than in one of the tourist resorts. Everywhere I went, would get stopped by locals and asked how my day was going, where I was going, what I was up to.

Shamefully my tourist-shields were at maximum after experiences in Morocco/Ethiopia and similar, and many people I ignored and kept walking as fast as I could.

Eventually I found myself in a conversation I couldn't easily escape from and I realised... they're just being friendly. They were all just being friendly. I spoke to dozens afterwards and had nice little chats, with no motives, no scams, no sales, no brothers-uncle's shop that I must visit.

(I did get scammed in the taxi though, by someone who didn't make conversation :) )


Never figured out how to get the buildx cache to actually work reliably on ARM OS X. Horrible if you have to build x86 images regularly.


Good luck getting ECMWF ensemble data for free.


That cost makes absolutely no sense. It takes one single day for a couple of people to install solar and batteries on a residential house.


Baumol's Cost Disease at work, I guess.


It's a third or fourth of the price in Australia with equivalent labor costs.

It's mostly unnecessary red tape and a broken market that cause the differences.


Australian labor costs are significantly lower than the US, as are labor costs in most countries. Americans are paid pretty well.


Solar installer costs are broadly comparable as Australians are better qualified and even if they weren't comparable the fraction of the cost isn't enough to explain the total difference.

There's various studies comparing the two countries, Tesla did one and found various technical approach changes and permitting reforms. It suggests labor is 7% of the cost in the US. Soft costs around acquisition, sales and marketing can be 18%.


Your expectations are wild. Most software engineers could not write a game boy emulator - and now you need zero programming skills whatsoever to write one.


Ctrl-C + Ctrl-V. There. Done!


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

Search: