←back to thread

We're Joining OpenAI

(www.alexcodes.app)
192 points liurenju | 8 comments | 03 Sep 25 18:38 UTC | HN request time: 0s | source | bottom
Show context
CyberMacGyver ◴[03 Sep 25 20:45 UTC] No.45120228[source]▶
At this rate it’s better to start a company and get aquihired vs applying and getting hired.

Seems like OpenAI speed ran through the Facebook phase and are out of ideas

replies(5): >>45120368 #>>45120639 #>>45123216 #>>45124631 #>>45125345 #
kridsdale3 ◴[03 Sep 25 21:00 UTC] No.45120368[source]▶
Au contraire, based on their recent hires, they're just beginning their Facebook phase. Expect to see a lot of ads 2 years from now, and expect to see a LOT of money being made by the company a year after that.

If they can maintain runway until then.

replies(13): >>45120439 #>>45120462 #>>45120737 #>>45120923 #>>45121451 #>>45123833 #>>45124176 #>>45124207 #>>45124568 #>>45125196 #>>45127156 #>>45128868 #>>45129566 #
1. simianwords ◴[04 Sep 25 06:24 UTC] No.45124176[source]▶
I predict this won’t be the case. The switching cost to a non Ad LLM is too low for me to be bothered with an ad based LLM.

OpenAI must justify ads when its competitors are not sending ads. Why would they? OpenAI models must be so good that it should be worth dealing with ads compared to say Claude or DeepSeek.

replies(4): >>45124812 #>>45125322 #>>45125776 #>>45126265 #
2. rat9988 ◴[04 Sep 25 07:59 UTC] No.45124812[source]▶
> The switching cost to a non Ad LLM is too low for me to be bothered with an ad based LLM.

Provided one with sufficient quality exists.

replies(1): >>45125222 #
3. simianwords ◴[04 Sep 25 09:13 UTC] No.45125222[source]▶
Yes so what amount of quality are you willing to give up so that you won't be bombarded with ads?

Or, what extra money would you pay OpenAI to get a non ad powered LLM model?

replies(2): >>45125457 #>>45125680 #
4. pjc50 ◴[04 Sep 25 09:29 UTC] No.45125322[source]▶
That requires you to be able to notice the ads.
5. mattmaroon ◴[04 Sep 25 09:53 UTC] No.45125457{3}[source]▶
The thing is, ads don’t pay that much. Facebook’s ad revenue works out to about $50/user per year . I probably don’t like FB enough to pay $5/mo for it (and even if I did, all the people I interact with that make it useful wouldn’t) but something that’s a personal assistant? Easy.

It’s not hard to imagine ChatGPT just charging.

6. andy99 ◴[04 Sep 25 10:34 UTC] No.45125680{3}[source]▶
I suspect that ad powered LLMs will also dramatically sacrifice quality- they are a cost center suddenly, no reason to run a model with 10s or 100s of billions of parameters when some 500m thing will provide a minimally plausible scaffold for ad delivery.

Like everything else, most people won't notice a difference and prioritize free over paid.

7. solarkraft ◴[04 Sep 25 10:49 UTC] No.45125776[source]▶
I predict it will be sneaky enough and the ChatGPT brand is strong enough to take it. Your switching cost is low - but you’re likely a tech enthusiast who likes researching and trying alternatives. Most people aren’t like that. The one competitor with anywhere close consumer recognition is Gemini and I’m sure Google would love to push ads too. Maybe not yet as they’re in the reputation building phase, but if it makes money I’m fairly sure everyone will end up doing it.
8. petralithic ◴[04 Sep 25 11:55 UTC] No.45126265[source]▶
Unless they all have ads, which is a good way to justify spending all that capex on training.