Most of the time people who complain about being censored are saying something very pedestrian that we've heard over and over again. (Like "A trans woman is not a woman and it is so terrible that nobody is saying that")
I was thinking of making a Twitter clone called "Me too" where you could press "Me too" whenever you wanted to signify your membership in a tribe as it seems a lot of the people who think they are being censored are really out to do just that.
He published this immediately after the 9/11 attacks to promote lisp as a solution for similar terrorist attacks.
I don't know what the backlash was exactly but he apparently felt ashamed enough about voicing these suggestions that he de-listed the article. But he still cares enough about their insight to keep the article available as an isolated undiscoverable island of a URL. It shows that feeling a sense of shame (from public reaction) can be overridden by the power of personal conviction, even if the result is a silencing they don't agree with.
I'm sure this 9/11 experience was fresh in their minds when this follow up essay was written. It's the only actual example I know of from them on this topic, it's evident it personally touched their family.
> ...he apparently felt ashamed enough about voicing these suggestions that he de-listed...
Perhaps. Admitting that I don't know the backstory...I'd say that "he didn't want to spend even more of his life fighting the Internet Hate Machine(tm)" would be ample reason for pg to de-list the article.
The essay is about a parallel between computer security (secure coding) and airplane security. It mentions Lisp as an example of a managed language.
Graham observes that what is ineffective in airplane security is analogous to something he believes is ineffective in secure coding.
Airplane security consists of screening what goes into the airplane. The airplane doesn't have protected domains. The payload space of the air plane (passenger area) is connected to the control space (cockpit): one can pass from one to the other. If it was impossible to get into the cockpit, then airplane security wouldn't have to rely on screening as much.
Graham claims that this is like putting code and data together into the same space and then trying to validate the values, like array indices, rather than by somehow separating things so that even if an array overrun occurs, it cannot be exploitable.
Graham's point is not very well made, but he's certainly not promoting Lisp as as solution for hijacking; that's just silly. Even GPT4 can tell you that. See below.
Some problem's with Graham's argumentation are:
- managed languages don't actually separate code spaces from data spaces. If a Lisp or Python vector could be overrun, it could go into some object where it helps the attacker divert a control flow. Managed languages actually resemble the screening of passengers at the terminal gates more than Graham observes.
- hijackings aren't the only problem; making it impossible for passengers to enter the cockpit might prevent a hijacking, but not a suicidal bombing.
- a bad agent boarding an airplane is not like an invalid datum. They conceal an intent that is inaccessible to screening. Browser security would be a better analogy for airplane security because your browser invites "passengers" to have can have harmful intent, in the form of remote javascript. And, indeed, browsers follow a security model which is more or less the inaccessible cockpit in Graham's article.
I'd delete or revise the article if I were Graham.
No, Paul Graham is not advocating Lisp as a solution for airplane hijackings in his essay “Hijacking is Buffer Overflow” 1. Instead, he is using the analogy of a buffer overflow attack in computer programming to explain how hijackings occur. In a buffer overflow attack, someone gives a program much more data than it was expecting, causing it to overflow into the memory occupied by the program itself and allowing the attacker to run their own code. Similarly, in a hijacking, passengers overflow into the cockpit from the cabin, taking control of the plane 1.
Graham suggests that the most reliable way to prevent hijackings is to separate the cockpit from the cabin, just as garbage-collected languages like Perl and Lisp keep code and data in separate places, making them immune from buffer overflow attacks 1. However, he is not advocating for Lisp as a solution for airplane hijackings.
> Thirteen years later, that's my default plan. There’s just too much downside for me to get distracted with others’ opinions of my opinions. [1] It's not that I'm afraid of expressing my opinions. I just think, "Why bother?"
Is it really true that she's not afraid though? It's a perfectly natural & understandable reaction to be afraid of the internet mob shouting you down.
The reason I ask is bc we should be honest about the problem, so that we can solve it.
Perhaps the solution is to teach the good, sensitive people around us how to have thicker skin, so they don't have as strong of an emotional reaction to internet backlash. But, detaching yourself from ego is actually a really hard skill to teach & learn. It almost seems like a hopeless endeavor when you're talking about creative people who are already very busy.
Also, a microcosm of this that applies in our daily life is how we interact with our coworkers. Do we build a positive environment where we nurture people to ask questions, propose ideas, and enrich the conversations around us? Or do we try to one up each other, deride ideas, and ridicule those less knowledgeable than we are?
There’s truly something to be said about assuming good intent on everyone’s behalf. What JL describes here is what happens when that collapses.
The term is both considerably older than that, and has shifted meaning over the decades.
In the 1980s, "fake news" referred to pre-canned "news" segments created by commercial interests (or their trade organisations) as "video news releases" (VNR) or "audio news releases" (ANR).
That was the context in which the phrase "fake news" was being discussed as recently as 2007/8.
I was thinking of making a Twitter clone called "Me too" where you could press "Me too" whenever you wanted to signify your membership in a tribe as it seems a lot of the people who think they are being censored are really out to do just that.