@Chagsy im almost tempted to take the other side of that bet. Note AMZN is suing Perplexity to prevent its crawlers from entering the AMZN site; it will be an interesting journey to watch. plenty of water to go here, i think

I personally reckon we’re going to see a whole new set of companies emerge out of the AI wave. Some of today’s mega-cap tech names will probably get disrupted, and completely new players will take their place.

Elon Musk said on a recent Joe Rogan episode that he thinks future phones will just be a simple screen running an AI layer with no apps at all. Interesting idea… but I honestly can’t see apps disappearing completely. There’s too much legacy infrastructure, too many businesses still running on 80s/90s systems, and too many workflows that aren’t going to be magically replaced overnight.

If Amazon was going to be replaced by some LLM who is going to do inventory and fulfilment? LLM at the moment are relying on IP (for lack of a better word) theft. If Amazon cuts off access would that mean a disaggregating to multiple sites again?

Then how is the LLM going to make money, is everyone going to start paying for $500 a year for LLM (probably more) as it replaces an advertising model? They’re current being massively subsidised and it’s capital intensive. Not sure scale benefits will accrue unless the data centre owners take a massive bath on their investment.

Think lots of question. I’d think the current aggregators have the power but it can change like the balance of power being fought between spotify and the music labels now, with spotify in a much better position.

All so uncertain. Hard to put any special credence on anything Musk says seeing he has been wrong on so many things in the past. It is a plausible idea though.

Thanks for the thoughts @Solvetheriddle @Magneto and @Mujo

The pod discussed the mechanisms of how such a thing might come to pass. I should clarify that the conclusion was not that this was going to happen, just that it was a possibility. There were a lot of "ifs and buts", and some serious improvements in user operability required, but there could be pathway to this becoming real.

To try and answer the points raised.

AI crawlers would not need access to AMZN's platform. Brands are already re-designing their websites to improve search results of LLM's much like the SEO techniques of yesteryear to improve Google ranking. The LLM search for a product would thus take you directly to the products' website, bypassing AMZN.

As such, the company would directly fulfil the order. Currently, no one researches or searches for a vacuum cleaner by directly visiting the Dyson website, but if we use LLMs to research reviews or particular specifications we are interested, in the next step would be either links to the product or a follow up question by the LLM to take you to the store. Or indeed the cheapest place to buy that vacuum cleaner; be that the Dyson online store, Goodguys, Appliances direct etc. It could even be Amazon. The crucial difference is that the default search is no longer hopping onto AMZN and searching on their platform.

As to how LLMs make money, I have no greater insight than any of the other people who are wondering about this, and this was one of the counter-arguments discussed that resonated most with me. The response was that they would either have evolved into something useful enough that we are happy to pay for, or that they take a cut of the sale. The latter seems more plausible and would introduce a pathway to significant revenue growth for the makers of LLMs. There is already a working version of this in fashion retail: https://daydream.ing/ - they charge 20% of sale price as commission. Here's a Forbes article on the subject

But agree, many twists and turns to come.