Hacker News Reader

Comment by ausbah | Hacker News Reader

scariest stat was .1% of companies generate >60% of US GDP? seems false but?

That doesn't mean anything. You can accomplish that by having a very large denominator, which is possible if creating a business is trivially easy.

In the US, it's generally very easy to create a "business," which doesn't even need to be a separate legal entity.

I'd guess that a substantial portion of the companies in the denominator have no employees and are just entities for owning a piece of property, or for someone's occasional graphic design work, or whatever. They are legally businesses, but not businesses in the way that we commonly think, such as a local auto shop or retail store.

> You can accomplish that by having a very large denominator, which is possible if creating a business is trivially easy.

This is correct, but misleading. In the US, something like > 60% of GDP come from just the US companies in the Fortune 500.

Reverse also true, many of the big businesses have hundreds of legal entities but only are counted as a single big business

I once worked for a F100 sized healthcare firm that had a separate LLC for each physician it 'employed'. Quotes because the LLC invoiced the parent and then paid the physician, so they were technically an employee of the LLC, but it shielded the company from certain risks and had some tax reasoning and various other mumbo-jumbos

Quick google gives ballpark figure of 30 million businesses. Companies are something like 17-19 million, 0.1% would be simple divide by 1000.

So we get 30 000 or 17 000. Which is entire SnP 500. And probably even every publicly traded company.

Your forgot government, which is about one third in the US

Most US companies are actually tiny, with pretty low revenues. Per Copilot, 82% of businesses have no employees and generate an average annual revenue of just $44,000.

For every US Steel, there's a thousand potters throwing vases in their backyard shed.

Copilot is not a source. Find where it got that number or don't post. Seriously. It's actively harmful to cite a language model. Would you cite Autocorrect?

It's a tool like any other. Get used to it.

It's a tool, not a source of empirically valid data. No one's going to 'get used' to LLMs being cited as sources of factual information any more than they'd complacently accept someone's Monte Carlo simulation or a hypothetical thought experiment being cited for the same purpose.

Additionally, outputs from LLMs can vary by user or prompt, making the data unreliable without clear sourcing or context.

If you said you Google’d it, we’d still ask for a source.

The thing I find amusing is that if I hadn't mentioned Copilot, no one would have batted an eyelash!

If you claimed that your source was The Onion, you’d get similar responses. You can make a claim, with no source, and expect the traditional answers. When your source is a “known to produce crap” source, what would you expect to hear?

Reply on news.ycombinator.com