That wasn’t my takeaway from it at all. I think he’s quite precise in his definitions and the point he’s making about Rust, even if I found the prose a bit fatiguing due to the LLMisms.
Personally I think that the world needs more of your optimism. But what does your crystal ball say about energy costs and political stability?
It’s one thing to say that AI will help everyone create immersive games, but skyscrapers won’t be free unless energy is free. Do you also assume that AI will solve fusion?
What happens to the individuals who are not “strong” and how do they hold onto their remote work jobs? Why will the biomedical research technology you’re positing not be used to create biological weapons, or do you assume that AI also creates universal peace and harmony? If so, how does it do that while also preserving our ability to have our own ideas and disagree with each other?
If we want the great future you’re imagining, I think history teaches that we need to give at least as much attention to these questions as we do to making the technology work.
I think their work earns "theory" because it makes specific predictions both about how to make more effective prompt injection attacks and what activations you'd observe in the LLM during those attacks, and can also be plausibly extrapolated to suggest useful future research directions.
They were adopted by professionals long ago, and those highly tuned and validated proprietary models are going to kick the butt of the models that you have access to every day of the week.
Claude is already shockingly good at reverse engineering. Try it – it's really a step change. It has infinite patience which was always the limited resource in decompiling/deobfuscating most software.
I think the origin is in the phrase, "Will the dogs eat the dog food?" which was common VC-speak in the 90's and 00's, referencing dog food commercials that once ran on TV, and meaning something like "this has been made to sound great in an internal powerpoint presentation, but will customers actually like it?"
In 2015, Marc Andreessen memorably said of Mixpanel's success at product-led growth: "The dogs are fucking jumping through the screen door to eat the dog food. And he hasn’t done any marketing yet."
https://www.newyorker.com/magazine/2015/05/18/tomorrows-adva...
That then led to the idea of "eating your own dog food", because if even you won't eat it, what credibility do you have saying that the other dogs will?
It's not "the largest representable number" because you're not representing numbers in any rigorous sense. If I give you 64 bits, you can't tell me what number those bits represent (first, because the rules of the game are ambiguous - what if I give you 8 bytes that are a valid program in two different languages; and second, because even if you made the rules precise, you don't know which bitstrings correspond to programs that halt). And if I give you a number, you can't tell me which 64 bits represent that number or even if the number is representable, and that's true even for small numbers and even if I give you unbounded time.
It seems far more natural to say that you're representing programs rather than numbers. And you're asking, what is the largest finite output you can get from a program in today's programming languages that is 8 bytes or less. Which is also fun and interesting!
Heat is not by itself waste. It's what electricity turns into after it's done doing computer things. Efficiency is a separate question - how many computer things you got done per unit electricity turned into heat.
The GraphQL ecosystem has grown amazingly quickly over the last year. It's definitely not a single-vendor technology at this point. Check out this list of GraphQL libraries, tools, and implementations:
https://github.com/chentsulin/awesome-graphql
The majority of people who are using GraphQL are using an implementation from someone other than Facebook (on either the client or the server, or in many cases both).
(And for what it's worth, I did see a "Why not RDF" slide in one of Lee Byron's decks, and those of us at Meteor who are working on GraphQL are definitely aware of the RDF/SparQL roots. I think what's driving GraphQL's growth is, first, it addresses a very timely problem - fetching all of the data for a screen in a mobile app in a single round trip without coupling your backend to your UI - and second, the focus on tooling and developer experience which has been a weakness for SparQL.)
Our Meteor business is thriving and we beat our 2016 revenue goal by almost 40%, driven by strong Galaxy growth. We expect that to continue into 2017 based on what we heard from a survey of Meteor/Galaxy users that we did recently – they are using Meteor for mission critical apps and most of them plan to write more Meteor apps in 2017. In line with this, we are growing our Meteor open source team while also continuing to ramp up our work on Apollo.
The big difference between Meteor and Apollo is that Meteor is about new app development (specifically in JavaScript), and Apollo is about a data system that you can add to already-existing apps that are running in production at meaningful scale. There's a place for both of these things and a lot of overlap between them.
Meteor isn't going to take over all JavaScript development the way Rails took over Ruby development, at least not anytime soon. That's just not how the JavaScript ecosystem works. However, I think it will be the #1 full stack JavaScript framework for a long time to come and will continue to be a really great option for teams that want to build JavaScript apps quickly, especially apps that have a realtime or collaborative element.
There are a variety of theories of copyright infringement in the US: direct infringement, contributory infringement, vicarious infringement, and inducement to infringe. I don't remember exactly which would apply to the circumstance you're describing, but in general, if you're making money off of someone else's infringing activity, if you are tolerating infringement that you have ability to stop, or if you have a landing page that suggests that people use your product to infringe copyright, you're probably in hot water, even if you're not infringing copyright yourself.
I dropped out of MIT after a term (in 1999), raised venture capital for my first startup shortly after that, and went on to have a successful career.
I never recommend skipping college to anyone. If you have a unique opportunity that can't wait, you have a solid plan, AND you have unique skills, then it might make sense. Or maybe if you have a very strange combination of strengths and weaknesses such that you are mature enough to succeed on your own, but not capable of making school work for you. Or if you can't find a way to get to a school where you will be with a good set of intellectual peers.
I know quite a few people who have dropped out of or skipped college, and in all but one or two cases I think they're worse off for it, or at least, they spend many years struggling to replace or replicate the college experience.
There are three big things you have to replace if you skip college: The social experience (learning how to exist both socially and intellectually in a group of peers). The material (in some hypothetical sense you can learn it all on OpenCourseWare, but in practice, most people find this to be a serious grind). The intellectual discipline (learning to think clearly and maturely, filing down your rough edges).
Your best bet is to find a group of really smart people who will be absolutely merciless in instilling intellectual discipline in you (constantly challenging you be rigorous, to know your field, to fully back up your ideas), and that will also be your close friends and romantic interests, and then find a way to have lots of spare time to hunker down and work through OCW or your favorite MOOC or textbooks. I know several brilliant people who have created these situations for themselves, and I think they've all found that it takes a huge input of energy to even approximately replicate what is readily available at top colleges, even for people that are smart enough to breeze through or basically intuit/rederive the course material.
Logistically, dropping out will close some doors forever. People with degrees can switch fields later in life by going to grad school; it is harder to switch fields if you don't have a degree. Immigration situations are far more difficult without a degree. There are ways around the closed doors but you will have to fight very very hard for them and become the top in your field. On the other hand, you also have a huge advantage (at least if you skip college entirely) which is that you don't have student loans and that may significantly increase your freedom during a period of your life when freedom is critical. In my own life I think these logistical factors came out about even.
In the end, I think it was right for me to leave in '99 and dive into startups, but, like, I made that decision in Stockholm, at the Nobel prize ceremony, where the establishment sent me after winning the top prize at world science fair, so I had every possible advantage and it was still a very difficult road. Do it only if you really don't see a future for you at college.