Hacker News Reader

Comment by a_band | Hacker News Reader

"As for the world outside of MIT’s PR materials, it appears much the same as it was in 1969—just with faster computers and uglier cars."

I this sentiment is actually right, especially when you realize that Thiel has made it clear that he isn't all that impressed with the advances of the last 30 years. We've seen innovation in software but not many breakthroughs in physics and atoms. For example, the predictions of the 50s and 60s about how we would live (flying cars, greater health, space travel, cleaner, futuristic cities) seemed to hope for those breakthroughs. The reality of where innovation happened is much more depressing: pocket computers, information superabundance, and porn.

For example: there has yet to be a breakthrough in the last 50 years that could even come close the impact of antibiotics and nobody has set foot on the moon in five decades. But if you told someone that in the late 60s, I think they wouldn't believe it.

View on HN

"For example: there has yet to be a breakthrough in the last 50 years that could even come close the impact of antibiotics and nobody has set foot on the moon in five decades. But if you told someone that in the late 60s, I think they wouldn't believe it."

This shouldn't be that surprising when you think about the situation just a little bit. Antibiotics are important because they cover a wide variety of diseases that formerly were untreatable. This is the case because they all have the same cause: bacteria. Other classes of disease, such as cancer, aren't so simple. There isn't a single treatment that can reliably cure all forms of cancer and theoretically there really can't be. It's just a fundamentally harder problem. There's a lot a low hanging fruit in basic discoveries but then we hit a wall when we attempt to apply them to more difficult problems.

Back in the 50's (I think) there was a fascination with the possibility of weather prediction. It seemed reasonable at the time that because the weather was ultimately a deterministic system we could predict it months in advance just by figuring out the physics and running the correct calculations. Then we discovered chaos theory and "sensitive dependence on initial conditions" and we realized that weather prediction was a fundamentally much harder problem than we could have predicted. We simply can't ever measure the atmosphere accurately enough to accurately predict the weather more than a few weeks in advance and even then we have to take a probabilistic approach.

There's a similar deal with fast planes as mentioned in the article. Back in the 60s and 70s all the big aerospace engineers wanted to be working on supersonic airliners like the Concord. Decades later, that effort seems to have been a waste of time. It turns out that the physics of supersonic flight present challenges and the economics of intercontinental travel don't really cover those challenges. The changes that have been made to airliners over the last few decades have been instead in terms of efficiency and safety. Even in the military sector, which can throw tons of money at pie-in-the-sky ideas has run into this. The most populous strategic bomber in the US Airforce is still the B-52 which was introduced in the mid 50s. We do have more advanced weapons, but we can't afford them, and we can't afford them precisely because of the changes that make them more advanced.

If someone in the late 60s wouldn't believe that we'd not repeat the success of antibiotics or return to the moon, it would have been because of misconceptions about the nature of the problems involved in the 60s rather than because of some form of decadence that settled in after the fact.

It is a matter of fact that the reason we’re not flying commercially x5 as fast is NOT an engineering problem. Ditto for reusable rockets before Elon. Ditto for electric cars before Tesla. In those cases, it’s just complacency. And there are many such cases

I think your argument risks being an example of how Thiel might argue that we educate ourselves into complacency. The discoverers of antibiotics had no reason to believe there would exist a treatment to cover such a broad spectrum of diseases; it's only in retrospect that we understand the mechanisms within which we say "we have found the low hanging fruit". Fleming's mode of inquiry and curiosity led him to discover a new branch of tools, but the article contends that our focus on specialization has taught us to bury ourselves in complexity and accept feeling relatively helpless.

Most of these advances (besides flying cars, they're just dumb) are social advances, not neccessarily purely technical ones.

Our cities could be much more futuristic, but that's clearly not due to a lack of technology - it's due to a lack of clear political will. Technology regarding health could be better, but start at implementing a decent health care system that doesn't let people avoidably die because they can't avoid treatment (this is US-specific, but a bad situation made even worse by the president Thiel so vehemently supported). Space travel, again, is in a large part just a matter of funding (we could be at today's point 20 or 30 years ago).

I would argue that we've seen only minor evolutionary advances in software in some areas, while others have actually gone backward.

I've been coding since I was a kid in the 1980s. By the late 90s I was developing multithreaded stuff in C++ and Java 2 that didn't feel all that different from today, except that Go has replaced Java for the high level server side niche.

UI development has gone way, way backward, especially on desktop. Today's experience is fragmented and clunky and just about every UI layer is ludicrously bloated. Other than somewhat better aesthetics all this complexity comes with no benefit or in many cases regression. The state of UI is absolutely horrid.

Meanwhile the Internet has become a surveillance and spam dystopia, but that's more of a social and economic problem than a technical one.

The only software advances I can name are safe multithreaded bare metal coding (Rust), somewhat more adoption of aspects of functional programming, and better dependency management. It's barely anything at all compared with say 1970 to 1990 or even 1990 to 2000.

Reply on news.ycombinator.com