"Early model organisms"? Really???

One thing I always wondered about Apple specifically is how do they actually coordinate so many layers of integration -- materials, hardware of various kinds, manufacturing, software. How is that done on like day-to-day basis?

Will they train on all the code hosted there? :)

I think with regulations "right" is highly subjective and also time dependent. "Right" today is wrong tomorrow. "Right" for one group is wrong for another today and it might switch up tomorrow.

Maybe they did and maybe that wasn't enticing enough of a goal for a model? It is all just game of probabilities. One pathway didn't yield this particular outcome while another did.

Perhaps that's an unpopular take -- but I actually would like it go even further. I don't mind ads in fact good ones I find quite useful. Since you, my dear ChatGiPiTi know so much about me, just show me stuff that reflects what I value. In fact I wouldn't even mind paying for that. The problem now is less that of discovery and more that of curation, so if it helps me find good things (but "good" for my definition of good which in one context might be cheap and in other solid quality and in third ridiculously ornate for example) -- why not? If they do it anyway why not do it really really really well.

"just show me stuff that reflects what I value."

Ads don't show you what you value. They show you what companies think you will waste your money on. You need to wake up, my friend.

I don't really see much ads tbh. Brave blocks most of online crap and other stuff I guess my brain just learned to tune out because it is not valuable. That's my whole point -- I think the ever increasing ad spend and ever decreasing RoAS will (hopefully) yield some platforms to do a bit better job. Maybe OpenAI be one of them. Maybe not.

Consumers have agency. The best ad placement for a company is to a consumer who wants what they’re selling. It’s a symbiotic relationship.

Ads are often misleading though. You know, it "looks bigger in the picture" type of thing. If ads were honest, the hose hideaway my wife just ordered would have said "flimsy piece of plastic shit you will throw away after a year, but should get you through the summer".

I think it's easy to lie to consumers. Then, how much agency do they have?

The best ad placement for a consumer is one for the best product of the sort they want.

Taking money to decide which of the options to show is in conflict with the goal of showing the most appropriate.

How do you know that only a consumer that wants what an advertiser is selling would receive an ad for that product?

Certainly we know the inverse, what of it?

it turns out "consumer agency" is ineffective and that people are actually very susceptible to having thoughts beamed into their heads

Seriously? 99% of the ads are trying to sell stuff you dont need at all. But they want to sell that to you. Why would you rely on an ad to realize what you NEED and then buy it?

People don’t only buy what they NEED?

I have been generally involved in the advertising industry for over 20 years, initially as my primary profession and now more on the side. But what I've seen in that time is that advertising provides absolutely zero positives to the consumer.

Well.. doesn't it pay for let's say Superbowl? Or makes NBA possible? Didn't it enable early internet to mature and become internet as we know now? If you're talking about just the first order affects, perhaps I see your point. But broadly speaking, advertising fuels a gigantic part of modern society. Remove it, and who knows what we will find ourselves in.

Look, I am also not convinced that modern society, or at least the economic aspects of it, are providing a net positive for the average consumer either. Saying advertising enabling these high-cost ventures being a good thing makes me immediately question why these ventures are high-cost to begin with. Yes, advertising enables the superbowl. But the superbowl exists in order to sell advertising

This is a deep kettle of fish to be opening, but I don't believe that the world we have constructed around us is necessarily good for us.

kind of a bad example, but take google, search, youtube, Android. It's an extreme position to argue that those products exist to serve advertising, they are clearly useful to end users.

I don't think it is. Google in particular is just an advertising company. Youtube would not exist if it did not serve ads.

And ultimately, while yes your examples are modern services that provide us convenience and entertainment, at a deeply human level I still argue that they are not providing net positives to us. Bread and circuses.

I agree with the sentiment that it might be net negative.

But on the other hand I'm almost certain I wouldn't have pushed through and gotten a degree without the internet (3b1b and Khan Academy in particular). I don't know what the world would look like without Google so almost impossible to argue, but I would definitely classify (ex. youtube) as "clearly useful".

Just gross positives is the claim.

Ad is the negative, and the services are the positive.

how is enabling ever bigger and more expensive to produce forms of entertainment making modern society better? remove all that and we won't lose anything. we'll always find a way to entertain ourselves. likely, with more culturally beneficial and social forms of entertainment.

This part of your comment, “just show me stuff that reflects what I value,” makes me think about the political manipulation used in social networks. It’s scary to learn about priming and how it affects our judgment; now imagine a politically (or religiously) charged ad in the middle of your learning session with ChatGPT that leads you to certain conclusions. Doesn’t feel like a good thing.

You know that it won’t be doing that, the tech for that either just isn’t there yet, or is more costly than your average purchase.

Many companies don't have good enough products, so must advertise to get any sales.

I am generally against the idea. I have seen it argued that advertising can provide access for those who could not otherwise afford it.

I am not convinced,

I would like all AI be required to state who they are acting in service of (with the possibly of declining to answer).

AI should act, To the best of their ability in the interests of who they are in service of.

It should be illegal to deliberately configure an AI to act against the interesta of whoever they are acting for.

But.. if you are running Chinese model in the US, what difference does it make? Isn't the whole "scare" (khm khm) with Kimis is that now I don't need Claude, cause I can run Kimi on my own hardware in my own datacenter and it's maybe not as good as Claude July edition but it's is as good as Claude January edition.

It doesn't need to be as good. You can route to the appropriate model and save so much money.

I have sonnet do the thinking, deepseek does all the tasks. I've massively reduced costs with this approach.

Sure, and I think that flexibility further undercuts his "frontier labs will be fine" take, which depends on top US labs having pricing power.