hoppp

830 karmajoined 2년 전

Submissions

Do you make more money now that you use AI for everything?

3 pointsby hoppp4개월 전2 comments

Zapier just had a supply chain attack

2 pointsby hoppp8개월 전1 comments

comments

2시간 전discuss

How do you make sure you are not violating terms and conditions if they forbid injecting third party javascript into the website explicitly?

Does your LLM ingest terms and conditions too?

In some cases in the past, finding and using auth tokens in the front end has resulted in hacking charges, in those cases alleged hacker found an api key in the front end and directly accessed the database, which resulted in a jail sentance.

How do you solve this issue that some companies may become hostile if injected code extracts auth tokens and accesses backend?

4시간 전discuss

Anthropic/OpenAI should have full control over who accesses their service since it's basically an API they provide.

Chinese models are great, nobody says otherwise but many of the publicly available ones are distilled and not trained from scratch.

Adding KYC would crack down on distilling because they use thousands of burner accounts to do it and now those accounts would need a face to verify.

8시간 전discuss

AI providers like openAI or Anthropic are already on the hook for OFAC compliance, so if any terrorist group is proven to access their APIs they can face huge fines.

The same game played out in the crypto space. Local models like local wallets can't be regulated.

Accessing foreign services can't be controlled.

But regulated companies are fully liable for any use of their tools by terrorist organizations.

This is not about the citizens, it's about the companies protecting themselves.

8시간 전discuss

The regulation is not for Westen Africa, it's for USA.

American companies could maybe keep OFAC regulations for AI?

If the fintech sector can do it, AI can too.

I'm pretty sure most of the western african terrorists are blocked from using coinbase, so why not chatgpt?

9시간 전discuss

If the impact of AI is comparable to the atomic bomb then the least they could do is try to filter out terrorists.

I think this is a very convincing argument to regulate the space more.

10시간 전discuss

AI services will need KYC soon?

10시간 전discuss

Their brains are the same, what differs is the environment.

Humanity's superpower is the ability to copy and mirror each other very effectively. It does not require advanced awareness or intelligence to do it. Most people copy others subconsciously!

Majority of people won't contribute anything technologically, but they sure as hell can copy.

10시간 전discuss

Just find the proteins involved then manufacture them with yeast. Easy Peasy Lemon Squeezy

11시간 전discuss

Seems like a pretty good article to sit down and type out the code and spend an afternoon learning.

13시간 전discuss

It's about the competitive edge. If a public model can do it, everyone has access to it so nobody has advantage.

If you want LLMs to be your advantage you can train your own, that's completely valid.

Let's say you want to have a company that runs inference 5% faster, if everyone can do it your business model is worthless.

14시간 전discuss

That's not always the case imho. Think about advances in running inference, any innovation will happen in the details. Higher layer can stack gpu's but the implementation can still be improved.

Often small technical changes like "making a service 5% faster" are worth millions for large companies. That's all implementation.

14시간 전discuss

If everyone is using LLMs then nobody has advantage

In that case, humans with superior skills who can write code become the advantage.

That is important for companies that compete with each other.

14시간 전discuss

One reason to write by hand is to have intellectual property, since it was ruled that AI generated code is in the public domain by default, so licensing is hard.

Another thing is competitive edge, if you use claude and your competitors use claude then nothing really gives you an edge. AI is a commodity, not competitive edge.

The competitive pressure should drive human work because it's unique.

Bun has an interesting blog post about how it was ported. It did cost a lot, much more than hiring people would cost outside USA.

Extensions like pgvector, TimescaleDB would probably need to be ported tho, not sure how much but there are footguns.

I would say it's an average coder when it comes to writing functions because it keeps using regex. It might pass a benchmark but doesn't pass the smell test.

If attention is all we need then what is better

Hours of human attention

or a few seconds of AI attention?

I am not just talking about writing the code but the brainstorming that goes into it.

There is often no spec, just the old code, copied with the old bugs and with new ones sprinkled on top.

I have an issue with the precision of generated code.

LLMs sometimes confidently leave things out or they will overbuild.

I use them all the time but mistakes happen. It's not exactly a scalpel, more like a sledge hammer.

What about extensions that are not sketchy? Lots of good ones out there.