At the very least the exchange has to be audited. Currently we have no idea whether the prices are a result of natural supply-demand dynamics or whether the exchange keeps artificially pumping the prices with lackluster demand
Or design errors in the algorithms doing the bidding!
There's serious nerd sniping potential in asking how best to construct an automatic bidder, especially with the speed and scale requirements in place. It's an incredibly deep problem, and I don't believe there is a single right answer.
> There's serious nerd sniping potential in asking how best to construct an automatic bidder, especially with the speed and scale requirements in place. It's an incredibly deep problem, and I don't believe there is a single right answer.
The problem is that it is unwise to trust an bidding algorithm designer whose incentives are aligned against yours. Google benefits from higher winning bids.
Key point is that it's a system that you shouldn't trust, not any individual algorithm designer or implementer. Bugs that cost Google money will be found and fixed really quickly; bugs that make Google money will linger and go unnoticed.
In the ad tech space the only winners are the people building and operating the adtech; everyone else is a sucker.
The only truly novel version of it which I have seen emerged from the Turkish hypercasual games space, where they managed to construct a giant audience everyone else thought was worthless, funnel them into their games, and then use the attention in the games to sell access to this apparently worthless audience.
Of course the audience actually were worthless, because all they were really interested in was new free hypercasual games, so the real suckers here were other devs that paid to access this audience but didn't have the adtech chops to make the most of it before the players moved on, and they funded their competition in the process.
Isn't the point that some bidders will not know which it is when they make the bid?
RTB bid requests have support for the normal auction based bids but also indicating possible private marketplaces which may come into play depending on the exchange and seller, meaning that the highest bidder will have indicated to the world their price, but not won the auction, so they may (erroneously) conclude next time to try bidding higher, leading to price distortions.
I suspect a large proportion of advertisers still believe the whole thing is a nice transparent fair auction process, and have no idea of how convoluted it has become.
They do know if the bid is part of a PMP. They can still place an Open Market bid if allowed but they should reduce their expectation of winning that auction, even if they're the highest bidder not because of skullduggery but because the publisher has a prior arrangement with a DSP or advertiser.
Valuing a bid is a complex and interesting task. Ever since Second Price auctions started dying out DSPs should have moved to essentially algorithmic trading. A price calculation depends on tens if not hundreds of factors that are evaluted on a per auction basis.
Advertisers have been demanding more transparency into where their money is going for quite some time now. If you're an advertiser and your DSP isn't giving you detailed reporting into the fees they're being charged then it's time to move DSP.
I mean game theory and equilibria are universal. I don’t see why the basic rules of civilization would not apply to any level of organism sophistication.
This is very typical in reinforcement learning. You just expand the state to include some more time periods. It definitely raises some academic eyebrows (since it’s not technically memory less), but hey if it works, it works
Ads was THE thing that our tech industry was exporting to the world. I was astonished with the people who thought that tech would be not affected by the tariffs.
The world is still consuming. Even if the US tariffs everyone else, they still trade amongst themselves. The real risk to our economy is that the rest of the world stops paying for our services and cultural exports.
Sometimes, one stands up to the bully even if its not the most convenient thing to do, something about morals and culture. Bully thinks this is some arab market with standalone greedy participants just like himself, while the game is anything but that.
Bipolar bully on top just generates random chaos, the best course is to cut oneself off as much as possible as quickly as possible, which is clearly happening behind the curtains
Either you will pay more for domestic producers or (more likely) you will not pay at all since either domestic production will not ramp up for niche things, or you will not be willing to pay hikes domestic prices for the same utility.
The counter argument is that now you can skip boilerplate code and focus on the overall design and the few points that brainpower is really needed.
The amount of visualizations that i have made after chat gpt was released has increased exponentially. I loath looking the documentation again and again to make a slightly non standard graph. Now all of the friction is gone! Graphs and visuals are everywhere in my code!
> focus on [...] the few points that brainpower is really needed
The person you're responding to is talking about it from an educational perspective though. If your fundamentals aren't solid, you won't know that exponentially smoothed reservoir sampling backed by a splay tree is optimal for your problem, and ChatGPT has no clue either. Trying things, struggling, and failing is crucial to efficient learning.
Not to mention, you need enough brain power or expertise to know when it's bullshitting you. Just today it was telling me that a packed array was better than my proposed solution, confidently explaining why, and not once saying anything correct. No prompt changes could fix it (whether restarting or replying), and anyone who tried to use less brainpower there would be up a creek when their solution sucked.
Mind you, I use LLMs a lot, including for code-adjacent tasks and occasionally for code itself. It's a neat tool. It has its place though, and it must be used correctly.
reply