Hacker News Reader

Comment by gretch | Hacker News Reader

I'm not the OP you replied to, but there's a scene from the movie the big short that really punched the understanding into my head:

Michael Burry: I may have been early but I'm not wrong

some exec: It's the same thing!

I believe that a lot of HN posts are probably "right", but they are so far in the future and so non-applicable to day-to-day life, that it's indistinguishable from wrong. For example, everyone predicting Google is going to end - yeah all great companies eventually come to an end - if you want to be insightful, you actually have to be on time with it.

View on HN

HN (and futurists in general) tend to significantly underweight inertia.

People hate change, and absent a catalyzing event or order of magnitude advantage, it will take a decade for usage patterns to shift.

E.g. Kagi over Google

Kagi could hypothetically be 25% superior in all searches.

Most people don't know and don't care. Search is the thing that pops up and they type into. You can even change that? To what? There are others?

Firefox had tabs.

Chrome had faster javascript. (Although moreso substantial free marketing on highly-visited webpages)

That's the sort of thing it actually takes to shift users.

> Most people don't know and don't care.

Yet. Time is the enemy of ad-supported business models.

> Time is the enemy of

Time is literally the enemy of everything. The universe will experience heat death. The sun burns out. Nations will fall. You’ll die of old age.

The assertion you made is actually the most milquetoast safest prediction that anyone can make about the ad business and it’s obvious to everyone.

The thing that actually matters is “when?”

"Man drinks like that, and doesn't eat, he is going to die."

"When?"

> Most people don't know and don't care. Search is the thing that pops up and they type into. You can even change that? To what? There are others?

So we’re all still using Yahoo and Altavista?

I think that's their entire point - it took many years for Altavista and Yahoo to die, and Yahoo isn't even dead yet in some markets (see: Japan).

Google eclipsed their capabilities >25 years ago and yet some of them are still kicking around. Inertia is a very strong thing.

Yahoo! Japan isn't related to anything outside of Japan: the brand was bought by SoftBank a while ago. Also, I've never heard of anyone actually using it for search, though maybe older people still do (the same people who use fax machines at home...). Everyone I've met uses Google.

No the internet population at time of Yahoo was X (say 400 million desktop PCs) about year 2000. Then google started to gain ground and from 2000-2010 internet population went from 400M to 4Bn - and even if every yahoo user from 2000 stuck with yahoo they were just 10% of users - and those users defaulted to google (partly actual defaults, partly fashion, partly here is how you do a search")

> E.g. Kagi over Google > Kagi could hypothetically be 25% superior in all searches.

The difference is the business model. Except for a small minority of people in tech and with disposable incomes, the majority aren't willing to pay for Search.

So if Kagi wants to increase their adoption they would need another business model. Advertising (in one form or another) is the easiest one.

I’m not sure the problem is the business model. Bing has the same business model as Google and has been trying hard to get some traction, even integrating ChatGPT, to no avail.

Google is a verb, it’s the default, it’s the internet. Two whole generations grew up with it being their portal to information.

There’s nothing Kagi can do to dethrone Google, unfortunately. Its only hope is to carve a niche for itself and cultivate it.

They could build a phone operating system and hardware ecosystem, make themselves the default on that, and then pay their competitors billions to also choose them as the default? /s

> Google is a verb, it’s the default

Hence, the lawsuit ;)

To wit: "The market can remain irrational longer than you can remain solvent"

Or in another case, there's a political scientist, whose name I don't currently remember, who's made fun of for predicting 9 out of the last 5 recessions. Sure, he may have predicted the ones that happened, but a broken clock is also right twice a day. Your predictions aren't worth much if you can't avoid false positives.

EDIT: [https://quoteinvestigator.com/2023/11/01/predict-nine/]

I’d say that some HNer predictions will inevitably turn out to be right, but you can’t necessarily tell which ones, at this point.

To the specific topic: Even if users predominantly switch to AI assistants for search (and there will always be search because users want to find stuff), the search will still be based on an index over third-party data, and the third parties will still be trying hard to optimize that data to be more prominent in applicable searches.

You want to be early, just not too early.

For trading, obviously if Michael was 2 days early, that may have not been enough time to arrange the contracts he did. 2 years early means paying premiums while explaining to investors why you are doing this weird thing that everyone else in the world disagrees with.

It reminds me of a MinuteEarth video that the sweet spot of disaster prediction is around 2 days before. Shorter and it's useless, longer and you can't really evacuate people for it.

Great point! I think it all depends were you are on the X-Axis of the Amara / Englebart law of technology prediction: https://www.youtube.com/watch?v=vuL0MGJMbp8

In this case, HN posts may be predicting the portion of the exponential curve further to the right, but in fact, we probably still have a long ways to go in the trough before we reach the point the lines cross.

Reply on news.ycombinator.com