Originally on HN: 

It’s always interesting to look at old predictions– there was that Newsweek article that predicted the Internet was never going to be a big deal [1] (amusing because now Newsweek is online-only), and there were all the negative comments on the launch of the iPhone. [2] Also I believe Drew Houston’s “Show HN: Dropbox” thread has a famous “Why would anybody need this, I can do the same with <complicated procedure>” [3].

It’s tempting to think, “Ah, people! So terrible at predicting things.” I think it’s interesting to think about why that is.

The main problem, I think, isn’t that people make wrong predictions altogether. It’s that it’s very hard to see how things will change and evolve over time, and how the ecosystem will change with it. The “One Laptop Per Child” idea [4] sounds a little dated and silly now.

I suppose if we just remember that progress is continuous rather than discrete, and that a lot of seeming limitations can be overcome with currently-unlikely innovations, then a lot of predictions will be forced to be a lot more precise.

Perhaps the Newsweek prediction might’ve been amended to, “In its present form, the Internet is unlikely to change the world.”

The problem with predictions is– things rarely stay in their present forms, and the world around them rarely stays the same, either.

I think pg addresses this in his most recent essay, “What Microsoft is this the Altair Basic of?” [5] In his words– “they practically all seemed lame at first.”

So we have to learn to live in a world where our initially valid assessments of a thing might become rapidly invalid because of change. And this is where Nassim Taleb’s work about the problem of prediction [6] comes into play. Rather than trying to predict a particular outcome, it makes much more sense to focus on evaluating robustness and antifragility– “How will this thing respond to change? What are the potential upsides, what are the potential downsides? What will kill it? What will give it more utility?”

Even if the odds are really low that something might come around, if the payoff is high enough, it might be worth betting on. I think that’s the whole point of things like YC.









This post was published on July 31st 2016. Predictions added on at a later date will be marked with [year-month-day].


Hillary Clinton will probably win the US Presidential Elections. Was going to bet on Trump until reading more about her extensive political support network + witnessing the RNC and DNC.

[2016-Sep-06] Okay, the stuff with the coughing does seem like it might become an issue. What a crazy election. [2016-Nov-09] WRONG. Trump won.


[2016-Sep-01] Driverless cars are “becoming a thing”, the way Teslas were in 2016. Probably the main way people in Silicon Valley get around.

[2016-Nov-09] Donald Trump will surprise people with a surprisingly dignified Presidency. He will either avoid talking about “The Wall”, or he’ll say it’s a metaphor, or he’ll say we have more important things to do in the meantime.


[2017-03-07] Uber as a company will continue to grow despite the multiple instances of sexism and misogyny that surfaced in 2017.

[2017-03-07] Barack and Michelle Obama’s memoirs will more than break even, making over $80m in sales. Claim made on Bilahari’s FB post.

Some startling forms of human augmentation will be revealed. Disabled people using prosthetics with neural interfaces will start to seem like they have a better deal than regular folks, and regular folks will start wanting these modifications themselves.

We’ll also see interesting commercially available health-related implants that measure blood sugar, heart rate, etc. The fitbit will be inside the body.

Male birth control. Come on.

[2017-01-17] Travel app “Wander” will fail.

[2016-09-01] Everyone knows a gamer who plays VR/AR games.


Virtual and augmented reality will be as common and accessible as smartphones were in 2016. Virtual escapism will become a bit of a problem, but the pros will outweigh the cons. It will be useful for people to overcome PTSD, phobias, a great therapy/spiritual tool. We’ll enjoy playing games with the avatars of friends who aren’t in the same room.

Alternate forms of money (Bitcoin, etc) will become ubiquitous and easy to use. People will make all sorts of micropayments for things, tipping one another over social media.

Marijuana will become a mature industry in the USA, shedding much of its “stoner” image and becoming a medicinal thing that grandparents do.

The USA will still have regular mass shootings. Unarmed black people will still be killed pretty much for being black.

Driverless tech is good-to-go, but adoption is being held back because of legalities and regulations.


[2016-09-01] VR/AR games are getting popular/serious. Someone makes a million dollars in prize money on one of these games.


Driverless cars will become the primary mode of transport in all major global cities with population > +1m. (The tech was probably good-to-go by 2020, but the legal complications would’ve taken a while longer to wrangle. Driverless cars + Uber or Uber-type service will drive down the cost of getting from point A to point B so much that many people will not bother buying cars at all. A human driving a car will seem as archaic/quaint as a human operating a lift, and it might even be seen as a public health hazard.

[2016-09-01] Virtual reality / augmented reality type deals would’ve taken off in some way by now. People will be using them at home and at work,


We’ll realize how cruel/barbaric it was that we didn’t make LSD, MDMA, psycobilin, etc accessible to people in a healthy, safe environment.

Euthanasia will be legalized in Singapore, Japan and similar homogenous-ish countries with aging populations.

Alternatives to the conventional University education track will become socially acceptable.




I will be dead.

Leave a Reply