I agree there's a real bias issue, but that is consistent through out any large company - e.g., Google, Facebook, Amazon, etc have sponsored results
Now that they are a commercial entity, you are right. If they had remained on their original mission path though I suspect this wouldn't be the no-brainer it is now.
If the latter was true, Google would take its tremendous cash balance and earnings and reinvest into what is necessary to reach AGI right now.
OAI has a funding problem. Google in comparison most certainly does not. Google could go raise funding for it right now too.
The real issue is the lack of confidence behind the supposed theory that more aggressive reinvestment yields AGI.
So if your users are spending a lot of time on a tasks, why not make it more delightful?
Of course, you need to make sure you don’t allow accidental purchases, that would be the way you destroy trust. But assuming a clear intent to purchase is established, then I think this will be well-received.
So I'm afraid of a steep enshittification of this use.
I think this is a blind spot for this community, personally. Like, it’s right to care about this stuff, but I think we are wrong to think other people care too.
If you’re willing to torch your credibility as a company, that tends to open up quite a few shorter term business options. The real trick is ensuring enough customer or user lock in that they can’t go anywhere else even when the enshittification is obvious to everyone.
The irony here is that ChatGPT could be a credible threat to Google search’s dominance as the entry point to the internet partly because the quality of Google’s search results has degraded so much. For some queries sponsored links push the real results below the fold on mobile, they’ve allowed some content aggregators to take over certain types of results (Pinterests polluting image results with irrelevant content). But that doesn’t matter while you make gobs of money. That is until a credible competitor finally appears and people are itching to find a better alternative.
That's only so in our little cynical skeptical contrarian hacker bubble. For most people, it's an appreciated convenience.
They might fall for ads. In the same way I smoked for years and hated smoking.
I think "people want frictionless ways to purchase products" is a PM pipe dream more than a description of reality.
I get what example you are referring to, but there are degrees here. For example the Buy Now flow really is handy; and I find I favor merchants that let me pay by scanning some kinda QR code from Apple Pay or Venmo. I definitely don't miss the friction of having to go dig out my credit card, mistype the cc#, type the wrong cvc if Amex, repeat the purchase after getting declined once and responding to a fraud text, etc.
How does Alexa ever compare with the rich experience of interacting with a store through the various senses? Its typical that technologists tend to come up with this stuff and what happens in reality is wildly off compared to what was expected.
Buying stuff means spending money. it turns out most people don't have a lot of money (something that Mr Altman would never be able to understand given his privileged background) so they want to see and experience the transactions that take place. Same reason why this agent nonsense is not going to work from an economic stand point.
That's a "fun" thought experiment if they did do that, because the advertising model inference will need to run cheaply enough compared to the conversion rate/margin on clicks. I suspect it will be really hard to beat Google Ads on cost if you have to run inference on all chat output for each ad placement. It could put ChatGPT into a higher end/higher cost advertising platform.
Sure they could. This notion that an unscrupulous revenue stream is justified if it pays well enough smacks of "Just following orders!"
"It happens that every man in a bank hates what the bank does, and yet the bank does it. The bank is something more than men, I tell you. It's the monster. Men made it, but they can't control it.” - Grapes of Wrath
At some point we have to stop this madness.
This implies antitrust has been sufficiently enforced such that there are viable alternatives in all markets.
That's not currently true.
The defaulting to negativity will really eat some communities up from the inside.
A voice assistant doesn't give you that option to review, but maybe it'd work for ordering fast food. A small chat window could grow to work for simple purchases like takeaway food or small hardware, etc.
OpenAI is projected to generate $12-14 billion in yearly revenue in 2025 (annualized from a single month), but expect to lose around 8 billion USD, implying the margins are negative.
OpenAI has raised a total of ~$60 billion.
I think they need to show investors a huge and growing cashflow to keep the show going.
Of course, they can't just retreat to selling their basic services since some other company would train and produce a marginally better model.
So it's a paradoxical situation. They're moving in contradictory directions - both to become a thing so valuable they'd only need to sell subscriptions and towards a mote if they don't reach that "AGI" thing. No reason being flexible would displease their shareholders but there are many other questions to answer here (who gets AGI raptures, who gets the Skynet/Terminator treatment, who decides, etc).
Im not really sure where Altman is going. As time goes on, it seems the walls are closing in and he's just throwing all he can to keep the hype alive.
You cant escape fundamentals forever, I dont care who you are.
They have hundreds of millions of users in total (free tier included), with around 10-15 million paying users.