Don’t forget the unbounded liability of AGI: your job disappears before you can accumulate the capital you need. On the bright side, you’re retired. Really simplifies the spreadsheet.
Don't forget the unbounded liability of apocalypse: society implodes and your money is worthless.
On the bright side, you might have invested in Nuka Cola.
My hot take is that open models don't really save you money and introduce more router complexity and security risk (because you're now sending your company data through more less trustworthy providers). Look at cost per task not cost per token and the pareto curve is largely owned by closed models.
Just use Fable 5.1/Opus max for the hardest problems, GPT Sol high as your workhorse, and maybe terra for async batch stuff you don't really care about. Gemini 3.8 High also looks pretty good and is quite fast if you're already a GCP shop. You can basically benefit from open models without using them because they force the frontier models to be cheaper.
We run Gemini fast for random end user queries for general staff.
We have our own on-premise inference server (quad MI300A) that runs Kimi 2.8 extremely well and we transitioned all heavy work to it since it's basically instantaneous for the whole team. It's a good enough solution and we will hit break even before the end of the year already.
Not everyone needs frontier models and availability is frequently much more important than a lot of companies realize.
Who gets to define the market? Does the "market" consist of only advertising technology platforms, or everything involved in online advertising, or all advertising through every possible channel worldwide?
Because bad behavior tends to scale with company size, not because smaller companies are more virtuous, but because they just have less power to do harm than bigger companies.
Imagine that there were only two companies, and they each had 50% market share in every market that exists. Would you expect either to be well behaved?
You may be surprised to learn that the US corporate tax (yes, that one, the main one we already have and everyone pays) was originally introduced in 1909 specifically as an antitrust measure, to clip the wings of US Steel and Standard Oil.
Of course the best laid taxes always get abused or turn into something else entirely.
There is a reference to this in For Whom the Bell Tolls:
“ Robert Jordan, wiping out the stew bowl with bread, explained how the income tax and inheritance tax worked. 'But the big estates remain. Also, there are taxes on the land,' he said.
'But surely the big proprietors and the rich will make a revolution against such taxes. Such taxes appear to me to be revolutionary. They will revolt against the government when they see that they are threatened, exactly as the fascists have done here,' Primitivo said.”
Seems like lead-generation for Vanta or other agentic-security tools? I tried to signup to take a look at the MCP and am forced to "Become a contributor" and supply a bunch of details to continue.
I like 1-2 days in-office if the team is actually local but the problem is that companies can't decide what they want (in-office vs cheaper LCOL labor) and it ends up being the worst of both worlds where you drive into the office just to sit alone in conference rooms on video calls with people in other offices.
It's not just LCOL options. For even modestly large companies, there are probably teams scattered around the world and you're having teleconferences at many hours of the day and night.
I call it 'agent speak' and it's pretty exhausting. Another pattern is it'll emit dense enumerations of a possible answer and then emit dense enumerations of what isn't the answer. It's still helpful to help rough out/frame a document but then you have to go in and edit it pretty heavily. "Make it shorter", "Make it more concise" doesn't seem help all that much because it still keeps a lot of the same patterns.
Why do you expect an AI developed in the same economic system would optimize for your preferences rather than their revenue? I think this only works if people can run, fine-tune and update their own models -- otherwise the giant AI vendors have every incentive to nudge you towards whichever retailer is bidding most on your Sponsored Responses, right?
I think the only difference will be that instead of advertising to target a query, they'll eventually need to target a vector and its neighborhood, and there will be some greater opacity facing advertisers trying to match "middle-income US users shopping for airfryers who have product height limits posed by kitchen upper cabinets" or something.
Yeah I truly do not understand some people’s logic. Like the ones talking about some post scarcity world where universal income will offset AI related job loss. So the people in power, the ones who have constantly tried to cut back government spending, are now going to flip and support universal income? Right.
Well same people had expected self-service restaurants, self-pay terminals will cause no tip. And even more fun future robotic servers can just work without expecting/asking for tips.
Or maybe I am wrong and "This time its really different".
I'd believe that if this article wasn't an obvious counter example to this.
People go where the mindshare is. It doesn't matter if its ad-ridden drivel and slop, it just needs to be popular.
The number of people who use OpenAI's products because "ai == chatgtp [sic]" in their minds is immense. Just like how to a ton of people, "online shopping == amazon".
I'm not arguing Amazon or OpenAI will always have this role, but while they have that the benefits of this position is immense.
and what stops the amazons of the world from paying the ai provider, and getting your ai to buy from them instead of the best deal? How are you going to check and verify you got the best deal?
In the recent past, there was a browser extension called "Honey", which purported to find and apply the best coupon code for you when you online shopped. Turned out they were actually (amongst other things), in cahoots with the shops by ensuring that you get only a modest discount, rather than the best available coupons (and this was paid for by the shop).
I have no doubt this will work out even less well than your typical SEO marketing spam site in the offering the best choice for you. AIs recommendations will be those of the highest bidder. Which will also be the lowest quality product they can get away with.
I don't really understand the value -add, so far "smart" routing currently produces worse results at higher cost and static routing can be done yourself trivially. Plus the risk of sending the most sensitive data you have to a third party.
Enterprises don't just send their data to any company with GPUs - they need security reviews, contacts, etc. They're assuredly using BYOK - not the marketplace.
reply