20 chunks · 17 episodes
...mizing for, so there isn't a ton of variance. We didn't mind this too much with Google Search. Google went out of its way to not have strong beliefs that colored the output. Anthropic is all about having strong moral beliefs.
If you showed a Google search engineer from 15 years ago today's results page they'd faint. Not only do many queries have AI-generated results. But also, every result above the fo...
Google Search for a long time was surprisingly simple internally. 90's era tech that stayed powerful for multiple decades. That's because if you have an emergent, ...
... what you do, you're unlikely to get hurt, and you'll likely get a good result. Google Search had this characteristic, and chatbots do too. The expectation for how likely a given input is to give good enough results is a prior that configures ...
...at via anonymous aggregation to produce crowdsourced intelligence. For example, Google Search's ranking is largely powered by the clickstream and the querystream. As Tim O'Reilly has said, data is like sand. Not useful in small quantities, but...
Google Search as a centralized service felt somewhat less scary to me than ChatGPT. But why? Both could have outsize impact from just a bit of bias given their sca...
Google Search started off as being most useful in the long tail of use cases. For all the things where the head of portals and manually-curated directories didn't ...
...nd tries are very hard to get to the bar of viability and then improve. Compare Google Search and (spoken) Google Assistant interactions. With Google Search, as long as the answer is in the top 10, it's fine. Formulating a query is fast, and t...
...gravity well of incentives into competing for users' attention. Many years ago, Google Search and Ads had a wall between them. Not any more. The wall was not bulldozed by any instantaneous force. It slowly eroded in the sandstorm of A/B experi...
...orrected its error and said that it's brown M&Ms, which I then confirmed with a Google search. A great example of the gullibility problem!
...ywheel. The more any user uses it, the more the quality for all users improves. Google Search has this property. Notably, neither of these is true anymore for OpenAI. OpenAI is no longer the best model. OpenAI also doesn't seem to have much of...
... judgment is in the loop, always. But the boring / hard part is automated away. Google Search is fundamentally framed as "here are 10 options, you pick the one you want" Very forgiving UX modality, with a clean hill to climb of quality. 4-up U...
...uite large; you're out $100 and only realize it was wrong days later. If it's a Google Search, the downside is tiny; just immediately do the query again manually. Amazon's search results have famously become quite scammy and crappy, which mean...
Google Search was selecting over a sea of static content. But in AI, content can be hallucinated on demand, so maybe the Google-like position is the generator of t...
...e a clear starter use case for users so they can get a sense of what it can do. Google Search and ChatGPT are two open-ended tools. But they are both good enough at so many things, that no matter what you do as a first use case, you're likely ...
...ost every MidJourney user session presumably has many iteration cycles, whereas Google Search sessions presumably have 1.1 (or less) iteration cycles. It allows the user to be in the driver's seat, but without having to be proactive. It gives ...
Google Search is a meta-product. Its quality and usefulness isn't just a reflection of the effort invested by its builders. The quality and usefulness is also in t...
Google Search is a hyper bespoke yet mass market product. It was one-size-fits-all, but also perfectly bespoke to each user's needs. The user interacts with it, in...
...t in the system. But it is not a given that it has to work this way! Consider a Google Search result where one of the ten blue links is clearly not a good result. It's not nearly as viral, Google isn't vouching for it as strongly. Going even f...
...aces you see this self-hoisting quality: Generative Adversarial Networks (GANs) Google Search's quality coevolving with user's expectations. Dunbar's Social Brain hypothesis, where massive increases in intelligence were driven by each person t...