Our government condemns and even demolishes people's houses when they're dangerous or become a public nuisances because their owners fail to maintain them.
Most Americans agree with this policy and don't see it as infringing on their freedom at all. Our choice is to enforce basic standards or allow our cities to become horrible places to live, and the same is true of the internet.
So yeah, it seems pretty reasonable to give people repeated warnings, then slow or pause their internet until they take basic steps to stop their connection from being used to abuse others.
> Fable-level results at 1/3 the cost using open-weight models
But we get ~$2500/mo worth of Fable credits for $200/mo on Anthropic pan? I'm still confused why people (who don't have to use API billing) are chasing open weight models based on cost.
What I'm calling for is exactly how things are already supposed to work.
The US government should already be infiltrating hacker groups and identifying infected American computers. Internet providers should be informing customers that hackers have compromised their devices.
Characterizing this as "mutilating the internet" is ridiculous.
> What happens when "those files on a USB drive" turn into a zero day button that can harm millions of people?
This already happens and the bad actors will get everything anyway, so all you're doing is diminishing the ability to defend.
> What happens when they help authoritarian dictatorships build nuclear weapons?
The knowledge of nuclear weapon construction is readily available to all authoritarian dictators. Most of them either already have (or don't want) nuclear weapons to continue in power. What stops them is the logistics and the CIA.
> What happens when they teach average people the steps needed to make a bioweapon at home?
There are a million ways anyone can do terrible things in the world. Fortunately very few people choose to do terrible things.
Humanity discovered a technique for compressing the world's information into a few files that fit on a USB drive. It's just insane that some people thought they'd get to be the only ones who get to possess them.
> Ok but these proxies almost certainly reverse-tunnel. How do you prove a customer is hosting one?
Why do you think that? These are proxies, so they're making huge numbers of outbound connections to websites on behalf of the people operating them. They are the "exit nodes" in this setup.
You could probably just count the number of unique destination IPs they connect to each day. If the average residential user connects to 5,000, an infected machine is probably connecting to 50,000+.
But the simplest approach is to buy access to these illicit proxy services and use them to make requests to web servers you control. If you see your own unique request arrive from a residential IP, you've proven that connection is being used as a proxy.
Not sure how it would? Because the bad actors are remotely instructing infected computers/devices to make HTTP/TLS requests for them, so they appear totally normal to the other side.
Okay, I was being imprecise. These residential proxies aren't "open proxies" in the traditional sense, but they're usually "open" to anyone willing to pay a small amount of money to use them.
> Most ISP abuse reports are routinely ignored. They might as well be a dead letter box.
This is where regulation might play a role, or at least a change in attitude. Companies shouldn't be allowed to pollute the internet in this way when they can easily prevent it.
In terms of law enforcement, sure. But what we really need is competent white hat hackers doing battle with the black hats.
Even if no one goes to jail, the NSA could make residential proxies much harder to operate in the US simply by detecting them and reporting them to ISPs. It would be good for ISPs to then validate the reports properly, give warnings, etc.
It may actually be costing them money through higher peak usage (requiring more capacity). I'm not sure but either way the government should have a role if market incentives aren't sufficient.
There's no spying required. The NSA and ISPs can find open proxies through infiltration and report them to (e.g. [email protected]), then Comcast simply has to act on it robustly.
ISPs already deal with abuse reports like this, the system just isn't being operated comptently.
US residential proxies are the bane of the internet. They're the major source of social media manipulation and spam.
Services can dramatically reduce abuse by blocking entire IP ranges based on country of origin, organization, or type (hosting providers). But a company can't block US residential IPs if it would also cut off many of their real customers.
The US government (probably the NSA) should be cracking down hard on US residential proxy networks. They're a genuine national security threat, actual data/identity loss of American citizens, act as infra for foreign covert influence campaigns, botnets used in hacking/DoS attacks, etc.
Major US ISPs (Comcast, AT&T) should detect clearly suspicious activity coming from customer IPs and warn them to scan their computers, check their TV apps, and find whatever is turning their internet into a proxy. It's very bad for the customers too (slows things down, gets their IP banned, etc)
Slack exists largely because IRC was insufficient. It didn't natively support channel history, search, etc.
For AI agents to flourish, Slack has to either truly open its network with a protocol or eventually be replaced.
I'd like to see Slack embrace an AT protocol-based chat system, which then apps like Buzz could implement. Then users could log in with domain handles like @yourname.com, and agents could use handles like @agent1.yourname.com, all under their complete control.
I'm here to learn by engaging with other people's ideas.
Blog: https://jacob.gold
Bluesky: https://bsky.app/profile/jacob.gold
Email: [email protected]