I don't know if this is why they did it, but I think it is generally a mistake to go beyond this context size anyway.
I don't think people realize how much dumber the models get at larger contexts and how much more the token cost is. I never let claude get about 300k and in general I don't compact, I just divide up the work in chunks where I can fit it into 300k and I try to keep really clean and slim docs and modular code bases.
It's still kind of annoying, sometimes the larger context is useful for one time tasks, but I think if you are regularly going beyond 300k you are losing a lot and probably have a poorly designed code base.
It's hell to maintain for not much gain (for a use case like this at least). As much as it's become a meme, JS and web tech in general has become extremely portable and stable.
I also don't think Anthropic bought bun to make their TUI better. They could have forked it, they bought bun because it incidentally was excellent for the way agents prefer to work and they wanted to capture that audience.
I have something similar. I use Cloudflare tunnels too but also their CDN which is free (for now).
That's the only thing I haven't really been able to figure out how to do on my own. Back in the day, hosting a static site from my crappy DSL connection was basically no problem and most people who were accessing my site were probably in my timezone. Now with how big the web is and how many bots there are, I worry about the quality of self hosting without a CDN.
If anything these models should be compelled to be public since they have been trained off public data. What an absurd overreach to call this an attack.
It’s clear they are scapegoating national security and China at this point to build an anti-competitive moat.
I generally really like Anthropic’s work and models but stuff like this scares me for the future. We are positioning these companies to have too much power. The public’s life is getting worse while these companies consolidate power using data they stole from the public.
I used to agree with you but it's getting out of hand these companies have more control than most states and governments over our communication, privacy, money, and general lives but we don't get to vote in who is running them. I'd rather start shooting and ask questions later before we get consumed by these behemoths. We need another Teddy Roosevelt and another antitrust sweep up in my opinion.
I don't know if Amazon should have the right to terminate anyone's account to be honest. I feel like people are afraid to regulate anything these days but I mean some people's entire lives are dependent on these corporate accounts. You can't just have me build up a livelihood on Amazon or have my whole house running on Amazon smart products and then just terminate the account.
If there was more competition or these things were based on open standards you could easily switch over I might be more inclined to agree, but I think right now these companies have way way too much power.
The threat of a company deleting an account for many people would be detrimental to their lives. We are way way too dependent on corporations and in kind of tired of people being scared to regulate.
I mean why are power lines not locked up and buried underground secured locked steel cages?
Because some things work better with trust vs convoluted security.
I think this is something a lot of computer nerds don't get (myself included at one point). It's almost like if something can be accessed we are allowed to access it and it's the fault of the person securing it. But a lot of our society works on trust and I think we'd live in a much more difficult world if everything had to be secure enough to resist any attack.
If this thing was connected to the internet I get it, but you already need physical access to the meter why add another layer of security on top of that? If someone has wants to mess up your power and they have physical access there's plenty of ways they can do it without wireless communication.
Yeah I had this experience too. I figured at least there'd be static checking that would at least make sure we aren't going off spec but there isn't really. So you just have a spec that slowly becomes out of sync with the code until it's basically useless. Just seems like double work for almost no benefit.
Web development has trained me to Google (now often ChatGPT) everything even if I already know how to do it because it changes so much haha
It makes me look stupid in code interviews because I commit very little to memory but professionally it's helped me catch a lot of syntactical upgrades.
I've been using GPT4 to code and these explanations are somewhat unsatisfactory. I have seen it seemingly come up with novel solutions in a way that I can't describe in any other way than it is thinking. It's really difficult for me to imagine how such a seemingly simple predictive algorithm could lead to such complex solutions. I'm not sure even the people building these models really grasp it either.
We completely privatized payments without realizing it. Credit cards are no longer some extra benefit on top of your bank account, it's how basically almost all commerce is required now. Private companies collect basically a tax on almost all transactions in our economy.
I don't know how we let them get away with this. Facilitating commerce should 100% be a government operation and be free. To accept cash I pay paying the US government a fraction of a percent I'm taxes compared to credit card fees.
The fact that, on top of all this, they are essentially policing what can and cannot be sold and bought should be outrageous to people. It's a complete conflict of interest that companies can shut down payments to and from competitors to their investments.
People can say "oh just don't use or accept visa etc" but good luck staying in business or buying anything online.
We absolutely need some kind of government backed digital payment system in my opinion. Credit card rewards should also be illegal they are nothing but hidden fees passed to you by charging the business more.
I'd like to see something like torrenting for websites. Just a way that if a certain site is important to you, you could seed it yourself and 1. Help with the hosting costs and 2. Preserve it if the original hosts decides to take it down or change it.
Yeah maybe, but I think we also already have a problem with overconsumption of media though. I am not sure this is helping.
It seems inevitable and I don't think we can stop it, but I just am kind of worried about the collective mental health of humanity. What does a world look like where people have no jobs and even creative outlets are dominated by AI? Are people really just happy only consuming? What even is the point of humanity existing at that point?
> If the pursuit is enjoyable, it should continue to be enjoyable as a hobby, no?
I think for most people the enjoyable and fulfilling part of life is feeling useful or having some expression and connection through their work. There's definitely some people who can create in a vacuum with no witness and be fulfilled, but I think there's a deep need for human appreciation for most people.
> As for your bleak future… will that even come to pass? I don’t know. Maybe it depends on your notion of “surpass”, and what that looks like.
I don't know either, maybe it will be fine. Maybe this will pass like the transition from traditional to digital. But something about this feels different...like it's actually stealing the creative process rather than just a paradigm shift.
To quote another comment but "Instead of replacing crappy jobs and freeing up peoples time to enjoy their life, we’re actually automating enjoyable pursuits."
I think this isn't just a simple discussion on competition and copyright, I think it's a much larger question on humanity. It just seems like potentially a bleak future if enjoyable and creative pursuits are buried and even surpassed by automation.
Yeah I agree. I was generally pretty pro AI art and agree with a lot of the pro AI sentiments here on a logical basis still, but as the tech develops I drift more and more towards thinking this may be a bleak path for humanity.
> Instead of replacing crappy jobs and freeing up peoples time to enjoy their life, we’re actually automating enjoyable pursuits.
Yeah really hit the nail on the head here. I thought a lot of backlash against AI was due to workers not really reaping the benefits of automation and that's a solvable problem. But I've seen a lot of artists who are retired or don't need to work dive into despair over this still. It's taking their passion away, not just their job.
I don't really know how we could stop it though without doing some sweeping Dune-level "Thou shalt not make a machine in the likeness of the human mind" type laws.
Fan art is pretty much illegal or infringement actually it's just not really enforced by most companies. There are some caveats for fair use but generally most fan art could be successfully taken down if a company was motivated enough in my opinion. Nintendo is pretty notorious for this but it has rarely gone to court as most people are too scared to fight takedown requests.
I don't think people realize how much dumber the models get at larger contexts and how much more the token cost is. I never let claude get about 300k and in general I don't compact, I just divide up the work in chunks where I can fit it into 300k and I try to keep really clean and slim docs and modular code bases.
It's still kind of annoying, sometimes the larger context is useful for one time tasks, but I think if you are regularly going beyond 300k you are losing a lot and probably have a poorly designed code base.