Hacker News new | ask | show | jobs
by SilverElfin 60 days ago
Yep. They can make every mistake imaginable and not work as hard but still win. It’s the power of concentrated capital and monopolistic behavior and what people call “moats” but really is just an unfair advantage. Why should Google or Apple be allowed to copy everyone’s AI tech and just win because of distribution through Chrome or iPhones?

We need new antitrust laws and heavy taxes just on the megacorps worth $500B or more. And aggressive enforcement.

4 comments

You mean, the inventor of the transformer technology that made ChatGPT possible, is copying ChatGPT’s technology?
Gemini is a copy of ChatGPT. And ChatGPT was a product invented on top of many previous ideas. The fact that one paper among many was written at Google isn’t relevant to my point.

Google entered the competition in AI products late. And now they will use their power unfairly to try and make it win. When they bundle an AI Chatbot into their existing contracts for Google workspace, they are competing unfairly. When the Chrome browser steers you towards Google properties by default, they are competing unfairly. Etc. Those unfair monopolistic actions let them come into the market years late with a viable competitor to ChatGPT or other products.

And let’s not give them too much credit for transformers. A handful of researchers were paid by Google while they came up with that paper. Google didn’t really do anything to push for it and neither Google leaders nor shareholders cared much about it at the time. Not to mention, transformers themselves were just a continuation of other prior steps in ML, from what I’ve read.

Let's not give too much credit to Bell Labs. A handful of researchers were paid to develop transistors and...

That's exactly how fundamental research works.

Transformers is possibly the most significant advancement in machine learning since AlexNet.

Bundling products is valid but different critism.

google literally had two divisions doing ai research. It is (was) risk averse and had its hand forced by the runaway success of oai.
Not that I'm opposed to new laws, but just having enforcement of the laws we already have would go a long way to fixing the problems.

The problem is how to get to the point where there is enforcement.

It definitely isn't going to happen with Republicans in power, and it also isn't a sure thing with Democrats in power either.

Lina Khan was a good start for a bit there, but she certainly didn't have universal Dem support. Establishment Democrats are going to have to grow a spine and tell the Reid Hoffmanesque donor class to get fucked.

The current set of laws lead to the current situation in my opinion. Enforcement within the current laws means a court case that will take years and span multiple administrations, which gives it a lot of time to be killed. It doesn’t provide enough authority to immediately bring enforcement actions.
I will preface by saying that someone with Lina Khan is sorely needed; Big Tech and other monopolies have gotten way too Big and seriously need to be reined in.

That said, from all the informed takes I've seen, Lina Khan was seriously... flawed (putting it charitably) in her strategy and tactics. To the extent that some observers wondered if she was deliberately sabotaging the agency just to highlight the need for new, more effective laws. She did have a novel theory of consumer harm, but that requires new legislation to enforce. Instead the way she went about it -- including by flouting due process -- was extremely counter-productive.

That was a big reason she was neither very effective in her goals (other than creating a lot of noise) nor have high political support from any side.

Her lack of political support from certain factions on the Democratic side was obviously because the big donors involved in the VC world want the option to continue to unload the startups they've invested in off as acquisitions to google, microsoft, et al.

Nothing noble about that stance, that's continuing to feed the Big Tech monster.

They are very much part of the problem that needs to be solved and they didn't like that she was starting to push for the solutions.

If you look into how she ran the agency, there were a lot of parallels with how the current US adminstration is being run. (Ask any AI for an overview and see the parallels pop up.) Regardless of your political leaning, I think generally we agree that is not how government institutions should be run. Even if the donor class hadn't made any noise, the Democrats were right in not supporting her tactics.
There are many valid criticisms of Google, but copying AI tech isn't one of them.
Everything it’s building now - Gemini in its various forms, and all the other AI products - are copies of other products. If they weren’t Google but another startup with the same products, they would be irrelevant and ignored. It’s their capital and anticompetitive practices that let them get away with missteps that no one else can survive.
What AI tech did Google just copy?