The internet was a tough sell. What's a similar technology today? | Hacker News Reader

The internet was a tough sell. What's a similar technology today?

6 pointsSteveVeilStream7 comments

I ran into a fellow at a party yesterday evening that had played a role in setting up one of the first dial-up internet service providers in a rural town in Canada in the 90s. He talked about going down to San Jose to a conference where he saw the first graphical user interface for browsing the web. He also shared stories about travelling to remote communities and having a hard time convincing people that this internet thing was going to be useful. This was only 30 years ago. Sometimes it's easy to forget how much has changed and how non-obvious the internet seemed in the early days.

Of course, there have also been a lot of other technologies that were a tough sell and never took off. In hindsight, we assume we would have been able to tell the difference but we're probably wrong about how good our forecasting ability would have been.

What's a technology today that is seeing a lot of skepticism and limited uptake but will have a massive impact over the next 10-20 years?

One idea that comes to mind is pre-fab housing kits/components. People have been trying to make it work for a long time but it's never been able to achieve scale. I think it will finally reach a tipping point.

RISC-V.

Internet replaced (and far surpassed) proprietary corporate and government communication systems such as X.25, Compuserve, ATT.net, AOL.

Linux replaced proprietary OSes. It is the biggest mobile platform (Android) and the vast majority of servers, MacOS is a parallel and largely compatible Un*x, and Windows is slowly being assimilated by WSL.

RISC-V is going to replace proprietary instruction sets such as x86 and Arm. Never entirely, of course, just as Mac and Windows and IBM S/360 have never gone fully away. They just lose their relevance to most people, and especially new users.

I was using internet in 1989 when New Zealand universities got the first 9600 bps full time connection to Hawaii, and a local BBS I was using (Actrix) got a connection to the local university. I was then able to telnet to BIX in the USA instead of using crazy expensive X.25. Usenet groups and email updated in real time. It was amazing. Limited then, but clearly the future.

I was experimenting with MkLinux part time on my PowerPC Mac in 1996, and x86 Linux on a borrowed Pentium 60 a year before that -- I had a contract to connect an app on the Mac to a mysql database on Linux. I bought a machine, a Pentium Pro 200, specifically to run Linux full time at home, in 1997. Linux also was limited at the time, but clearly the future.

RISC-V is the same now, and I've believed so since I bought my first board, a HiFive1, in December 2016. RISC-V is already well established in embedded uses, taking significant market share from Arm and virtually killing off everything else. In the last two years RISC-V has started to appear in cheap (close to Raspberry Pi price) SBCs and in 2024 in low end laptops.

I was playing around with various internet access methods in the 90s, including a weird system with download via a satellite dish but outbound data via modem which I got in I think 1997 (IHUG SatNet). And playing around with Linux.

Both started to feel really serious when I got both a cable modem and Ubuntu in 2004.

I think RISC-V will hit that stage in the next 12-24 months.

I don't know that it will fundamentally reshape compute as strongly as the Internet reshaped communications, but RISC-V is such an appealing platform. It's so powerful given power demands and cost.

The question was "have a massive impact", which is not quite the same thing as "fundamental".

Though being the Last ISA would be pretty fundamental.

Unless they are really really bad, ISAs don't die because their design got old. x86 has proven that, over and over. As has been observed by a number of people, the only flaw that is really fatal is not having enough address bits. RISC-V is probably the only 64 bit ISA today with a built in plan for 128 bits.

The main reason ISAs die is because they are proprietary and have only one vendor that goes out of business, decides to concentrate on something else, gets too greedy, misses the Next Big Thing etc.

Same with Un*x's, actually.

Something owned by no one and everyone avoids all that, as long as it gains initial traction and has a governing person/body that keeps things sensible but can be bypassed/forked later if needed.

The internet did not take off slow and it wasn't a hard sell. Maybe to home users in rural Canada, a negligible fraction of customers. Internet usage exploded so fast that just a few years after it became available to most people in developed countries we had the dot-com bubble and bust (2000). Everyone and their mother got online in a few short years. Before the public got access the proto-internet had already grown and expanded globally.

I can't predict the future. If I had any insight I would invest today and not tell anyone about it here.

I disagree. Invest today and tell Everybody about it.

Not a hard sell; didn't grow slowly.

Anyway, I'd guess Neural Implants. The next thing for rad body-mod hardcore net-citizens. Then a fad for hipsters. Then everybody has it. That's the path I predict.

FIDO2 like auth. I'm not sure it will end up being FIDO or passkey, but it seems obvious to me the world does need phishing resistant, anonymous authentication that seamlessly supports multiple ID's, can only be used by the owner, isn't owned and controlled by our Tech Bro overloads, and does have a workable plan for what happens when the token is lost or damaged.

Like the internet, it will require everyone and every company on the planet to converge on the one protocol which probably will be FIDO2 or something very close to it. Like the early internet running over POTS (plain old telephone lines) we don't have the hardware yet.

Yet it seems like we are getting groping towards a solution. Surely it must arrive one day.

Reply on news.ycombinator.com