HNHacker News
TopNewBestAskShowJobs

safety1st

6,473 karma · joined March 23, 2018

submissionscomments
safety1st··on We replaced RAG with a virtual filesystem for our AI documentation assistant
I agree it makes no sense. The whole point of chunking is to handle large documents. If your chunking system fails because a document is too big, that seems like a pretty glaring omission. I just chalked it up to the tech being new and novel and therefore having more bugs/people not fully understanding how it worked/etc. It was a vendor and they never gave us more details.

Not all problems have to be solved. We just fell back to using older, more proven technology, started with the simplest implementation and iterated as needed, and the result was great.

safety1st··on We replaced RAG with a virtual filesystem for our AI documentation assistant
We were given a demo of a vector based approach, and it didn't work. They said our docs were too big and for some reason their chunking process was failing. So we ended up using a good old fashioned Elastic backend because that's what we know, and simply forwarding a few of these giant documents to the LLM verbatim along with the user's question. The results have been great, not a single complaint about accuracy, results are fast and cheap using OpenAI's micro models, Elastic is mature tech everyone understands so it's easy to maintain.

I think this turned out to be one of those lessons about premature optimization. It didn't need to be as complex as what people initially assumed. Perhaps with older models it would have been a different story.

safety1st··on Go hard on agents, not on your filesystem
We anthropomorphize these agents in every other way. Why aren't we using plain ol' unix user accounts to sandbox them?

They look a lot like daemons to me, they're a program that you want hanging around ready to respond, and maybe act autonomously through cron jobs are similar. You want to assign any number of permissions to them, you don't want them to have access to root or necessarily any of your personal files.

It seems like the permissions model broadly aligns with how we already handle a lot of server software (and potentially malicious people) on unix-based OSes. It is a battle-tested approach that the agent is unlikely to be able to "hack" its way out of. I mean we're not really seeing them go out onto the Internet and research new Linux CVEs.

Have them clone their own repos in their own home directory too, and let them party.

Openclaw almost gets there! It exposes a "gateway" which sure looks like a daemon to me. But then for some reason they want it to live under your user account with all your privileges and in a subfolder of your $HOME.

safety1st··on Our commitment to Windows quality
This is a good time to reflect on how business actually works (as opposed to how hopeful consumers wish it would work).

A business exists because its shareholders invest capital with the expectation of a return. As a result, nearly all businesses go through similar lifecycles. The stages are launch, growth, maturity, decline, and sometimes renewal. There is a lot of capital injected in the early stages and to capture market share the firm often produces the best product it can.

Once the market share is acquired, the business puts up moats if it's able, and then it enters the MATURITY phase. That's where the Windows business is. In the maturity phase a business focuses on TAKING PROFITS wherever it can find them. This includes but is not limited to cutting back on its investment in product, as much as it can. If it can cut budgets and quality and give that money to the shareholders it will. If it can inject ads into the product or resell your data it will.

The very purpose of a business is to reach maturity and then take profits.

That's capitalism. The investors provided the capital. In the end, they gets what they wants.

Now if a company leans into this dynamic as hard as Microsoft has, you should know what's coming. No one should be surprised - maybe they're scared of the Neo right now and there'll be a few years of reprieve, but they're a mature firm, they're in profit taking mode, and the goal in this phase is not to make Windows as great as possible, it's to squeeze as much money out of it as they can.

The next stage is decline -- where the squeeze gets so hard that the business actually collapses. All businesses fail sooner or later. Everything becomes lawyers and accountants slicing it up, selling it off, and sometimes it gets restructured and reborn, sometimes it doesn't. This can take years or it can take decades but it's basically a bumpy downhill road from maturity to that point. If you stick around at this point and keep using Windows, keep in mind that's what you opted into. There isn't really any other way. It's just business.

Intriguingly, free software in its more elemental forms doesn't appear to follow this lifecycle. It's not for profit and there are no investors to satisfy. Contributors who build the software do it mainly out of self-interest: they build what they want to use, and as a result they may come and go at any time. But the software remains there, and you are welcome to tinker with it, too.

safety1st··on John Carmack about open source and anti-AI activists
This cultural shift exists and it will intensify as long as consumer prices and cost of living continue to rise at the same time corporate profit margins do. This is a simple, easy link to make, pretty much everyone's now aware and has stopped buying the excuses. Consolidation and an increase in straight up, unpunished criminal monopoly and cartel activity within corporate America have given rise to this new culture. Luigi Mangione will not be the last of his kind.
safety1st··on Create value for others and don’t worry about the returns
I mean it's a title. Titles are under no obligation to condense the entire content of the article into one sentence. People who want to comment on the article should read it first, and then write in good faith.

The problem lies in the HN comments which have taken that title and interpreted it through the lens of unrelated political arguments: class warfare, anti-offshoring, etc. etc. I don't think any title would be immune from these people. They're just angry because the Internet has its hooks in their brain, and they're going to post about it.

His points are good and people would be wise to read the article and take them to heart. His key points are:

1) If you're a rent seeker, current trends will probably see you lose out to a bigger and more powerful rent seeker. He's probably right about that.

2) Creating more value than you consume is a great form of self-preservation, when you do this no one wants to get rid of you.

None of it's political. It's just good advice for life. I hereby forbid the masses from responding to these points with political rage bait.

HN has better moderation than a lot of places but from my vantage point the entire Internet is sinking into this garbage - we're more aware of the problem these days, at least, but everything and everywhere is more consumed by political hot takes than ever before.

If there was tech that forced commenters to read the article before they could comment on it - now THAT would be a valuable innovation!

safety1st··on /e/OS is a complete, fully “deGoogled” mobile ecosystem
The alternative they accept is traveling down to their office and handing them cash, no joke. Phone app or cash. No website, never has been one. No snail mail because they "modernized" and discontinued it some time ago.

But I'm OK because one of my banking apps has some method of reading my contract number from the disabled electricity company app, and telling me how much I should pay and then it fires off a payment to them. Even though I can no longer use the electricity app directly because I enabled USB debugging once, the banking app is somehow still able to pick up this info from it.

Of course, said banking app refuses to run on Graphene or any of these other Google Play-less OSes, and the bank doesn't respond to inquiries about that issue, multiple people have tried.

The other bank I use does respond, and says they'll never run on "alternative OSes" because "alternative OSes are too insecure." They don't respond to followups.

I'm just saying man. A lot of people think this stuff is trivially solved because there is an option available to them in their home country. You don't know how big and nuts this world of 8 billion people and 200 countries is. This stuff varies beyond imagination, sometimes for the much worse.

safety1st··on /e/OS is a complete, fully “deGoogled” mobile ecosystem
It is great that you have the right in your jurisdiction to do that. Where I am, they just shut off your power if you don't pay.

It's a big and hairy world out there. Having lived on three continents and traveled to some pretty wild places, I always get a kick out of seeing which rights people have and assume that the rest of the world also has.

safety1st··on /e/OS is a complete, fully “deGoogled” mobile ecosystem
I appreciate that there are people out there working on stuff like /e/OS, but the number one question I have when I learn about a mobile OS that isn't iOS or "Googled" Android is: will the banking and payment apps I need to operate in the modern world run on this OS?

A lot of people don't think this way because they haven't had any problems. But then one day it happens to you and you realize, ok, this is the one thing that matters - you're in a cashless store and the only way you can pay for your meal is to use Approved Apple or Approved Google operating systems.

Where I live, the app my electricity utility provides for viewing and paying my account DISABLES ITSELF FOREVER if you so much as enable USB debugging on your phone (even after you've disabled it again).

To their credit Graphene maintains a global database of which of these apps work and don't. They're the only ones I know of so a thousand upvotes to Graphene OS.

But for my banks, the records in that database are grim. They won't run on Graphene, and they don't respond to reports about it.

One of my banks just discontinued its web UI because "people don't use it anymore, they use the app only."

This is how they're going to get us, folks. This is how we're going to lose it all. Writing code alone will not solve this. It will require some kind of collective action to defend our liberties. Some parts of the world are already lost. So this situation will likely come to a jurisdiction near you eventually: to make a transaction you will need permission from Google, Apple, Visa, Mastercard, or it won't happen. Then that four company list will start to shrink.

safety1st··on I am directing the Department of War to designate Anthropic a supply-chain risk
My main takeaway from all of this is that Hegseth seems deeply unfit for his job. First there was the Signal leak and now this.

Look, Anthropic is not going to be designated a supply chain risk. 80% of the Fortune 500 have contracts with them. Probably a similar percentage of defense contractors. Amazon is a defense contractor for example. They'd have to remove Claude from their AWS offerings. Everyone running Claude on AWS, boom gone. The level of disruption to the US economy would be off the charts, and for what? Why? Because Hegseth had a bad day? Because he's a sore loser?

If he's decided he doesn't like the DoW's contract then he can cancel it, fine. To try and exact revenge on the best American frontier model along with 80% of the Fortune 500 in the process, to go out of his way to harm hundreds or perhaps thousands of American firms, defies all reason. This is behavior you would expect any adult would understand as petty and foolish, let alone one who's made it to the highest ranks of government.

So I think it's just not going to happen, Trump's statement on the matter notably didn't mention a supply chain risk designation. This suggests to me that Hegseth went off half cocked. The guy is a liability for Trump at this point, I'm guessing he won't last much longer.

safety1st··on A new California law says all operating systems need to have age verification
My first reaction is that this is an insanely bad law:

* The signal has to be made available to both apps and websites

* So if you dutifully input valid ages for your computer users, now any groomer with a website or an app can find out who's a kid and who isn't. You just put a target on your kid's back.

* A fair share of parents will realize this, and in order to protect their children, will willfully noncomply. So now we'll have a bunch of kids surfing the net with a flag saying they're an adult and it's okay to show them adult content.

* Some apps/websites will end up relying on this signal instead of some real age verification, which means that in places like porn sites where there's a decent argument for blocking access from kids, it'll get harder. Or your kid will get random porn ads on websites or something.

So basically unless this thing is thrown out by the courts, California lawmakers have just increased the number of kids who get groomed and the number of kids who get shown porn.

Mind boggling that something this bad passed.

safety1st··on Pi – A minimal terminal coding harness
I think there's room for both visions. Big Tech is generating more toxic sludge than ever, and yeah sure this is because they're greedy, but more precisely the root cause is how they lobbied Washington and our elected officials agreed to all kinds of pro-corporate, anti-human legislation. Like destroying our right to repair, like criminalizing "circumvention" measures in devices we own, like insane life-destroying penalties for copyright infringement, like looking the other way when Big Tech broke anti-trust laws, etc.

The Big Tech slop can only be fixed in one way, and actually it's really predictable and will work - we need to fix the laws so that they put the rights and flourishing of human beings first, not the rights and flourishing of Big Tech. We need to fix enforcement because there are so many times that these companies just break the law and they get convicted but they get off with a slap on the wrist. We need to legislate a dismantling of barriers to new entrants in the sectors they dominate. Competition for the consumer dollar is the only thing that can force them to be more honest. They need to see that their customers are leaving for something better, otherwise they'll never improve.

But our elected officials have crafted laws and an enforcement system which make 'something better' impossible (or at least highly uneconomical).

Parallel to this if open source projects can develop software which is easier for the user to change via a PR, they totally should. We can and should have the best of both worlds. We should have the big companies producing better "boxed" software. Plus we should have more flexibility to build, tweak and run whatever we want.

safety1st··on Facebook is cooked
Indeed. However, there is no universal definition for what offends people, and never will be. People are individuals who form their own opinions and those opinions are diverse.

Ergo if you start to moderate speech which is offensive from one point of view, it will inevitably be inoffensive to others, and you've now established that you're a publisher, not a platform, because you're making opinionated decisions about which content to publish and to whom. At that point the remedy lies in reclassifying said platform as a publisher, and revisiting how we regulate publishers.

They can be publishers. They can censor material they object to. That's fine. But they don't need special exemptions from the rules other publishers follow.

I think it's good to have publishers in the world who are opinionated. There are opinions I don't like and don't want to see very often. Where we get into trouble is when these publishers get classified as platforms by the law, claim to be politically neutral entities, and enjoy the various legal privileges assigned to platforms by Section 230 of the CDA. The purpose of that section was to encourage a nascent tech industry by assigning special privileges to the companies in it. That purpose is now obsolete, those companies are now behaving like publishers, and reform of our laws is necessary.

safety1st··on A16z partner says that the theory that we’ll vibe code everything is wrong
Just in case you weren't aware, Gsuite has a clone of Docusign built into it now.
safety1st··on Claws are now a new layer on top of LLM agents
In my opinion people are fixating a little too much over the automation part, maybe because most people don't have a lot of experience with delegation... I mean, a VP worth his salt isn't generally having critical emails drafted and sent on his behalf without his review. It happens with unimportant emails, but with the stuff that really impacts the business far less often, unless he has found someone really, really great

Give me a stack of email drafts first thing every morning that I can read, approve and send myself. It takes 30 seconds to actually send the email. The lion's share of the value is figuring out what to write and doing a good job at it. Which the LLMs are facilitating with research and suggestions, but have not been amazing at doing autonomously so far

safety1st··on Facebook is cooked
This is the first time I've ever heard somebody claim that section 230 exists to deter child predators.

That argument is of course nonsense. If the platform is aware of apparent violations including enticement, grooming etc. they are obligated to report this under federal statute, specifically 18 USC 2258A. Now if you think that statute doesn't go far enough then the right thing to do is amend it, or more broadly, establish stronger obligations on platforms to report evidence of criminal behavior to the authorities. Either way Section 230 is not needed for this purpose and deterring crime is not a justification for how it currently exists.

The final proof of how nonsensical this argument is, is that even if the intent you claim was true, it failed. Facebook and Instagram are the largest platforms for groomers online. Nazi and white supremacy content are everywhere on these websites as well. So clearly Section 230 didn't work for this purpose. Zuck was happy to open the Nazi floodgates on his platforms the moment a conservative President got elected. That was all it took.

The actual problem is that Meta is a lawless criminal entity. The mergers which created the modern Meta should have been blocked in the first place. When they weren't, Zuck figured he could go ahead and open the floodgates and become the largest enabler of CSAM, smut and fraud on earth. He was right. The United States government has become weak. It doesn't protect its people. It allows criminal perverts like the board of Meta and the rest of the Epstein class to prey on its people.

safety1st··on Anthropic officially bans using subscription auth for third party use
There's no reason that a sizeable portion of LLM usage can't and won't end up free/ad-sponsored. Cutting edge stuff for professional use will probably be monetized via subscription or API credits for a long time to come. But running an older and less resource intensive model works just fine for tasks like summarization. These models will just become another feature in a "free" product that people pay for by watching or clicking ads.

I imagine the split will look a lot like b2b vs b2c in other technologies, b2b customers tend to be willing to pay for tech when it offers a competitive advantage, reduces their operating costs etc. b2c customers mostly just guzzle free slop.

safety1st··on Mark Zuckerberg Lied to Congress. We Can't Trust His Testimony
When you pick apart what's actually going on in Meta's revenue pipeline it's hideous. Think about this and compare it to what the world was like say 30 years ago:

* There are literally thousands of IG profiles that are essentially softcore porn which serves as a lead gen device for an OnlyFans account. Meta promotes these profiles to its users heavily because sex sells. Meta profits from the engagement with the profile, OnlyFans profits from signups sent to it by Meta.

* This is one of the primary ways OnlyFans has grown its pornography business to $8B a year

* Once users sign up for OnlyFans a common mode of engagement is that a managerial company lies and pretends to be the porn actress, and texts with the user under fraudulent pretense as the user consumes porn

Now... what was the world like 30 years ago?

* You couldn't buy porn mags without showing ID, Internet porn not really a thing for most people yet

* Even softcore stuff was mostly relegated to late night Cinemax

* Far fewer women had body image disorders and mental health disorders

* Far fewer young men had ED

This stuff is evil, when you connect the dots, it's crime, evil, lies and perversion all lined up to make a small number of companies a staggering amount of money. Somehow government and industry are OK with this, I guess this is the world the Epstein class built for us so no surprise. I am not a religious guy, and I would hardly call myself a prude, but this all exists and is widespread because it enables profit and fraud and exploitation, and I find that disgusting. Zuck's a porn baron. He knows what's going on. The fucker's on the take.

If anything should be in the dictionary next to the word evil, it's the 2026 state of affairs

safety1st··on Discord will require a face scan or ID for full access next month
In order to make sure businesses aren't giving porn to teens, you can require they do meaningful age verification at the time they want to provide the porn. You can impose criminal penalties on a domestic business which doesn't do this, and other penalties on foreign businesses (such as locking them out of the payments network). You don't need to get 100%, even partial success will act as a deterrent. This is how the world worked before the Internet, you needed to show ID to buy porn, and public opinion is in favor of the world working this way again. Crucially, penalties on businesses (not consumers, and starting with the biggest ones) are the way you need to go because this is the only way this can feasibly be enforced.

The libertarian concerns around privacy, freedom of expression and surveillance are all valid, but they're downstream. We have hard evidence that porn damages sexual health and relationships, and it has basically zero value to society; it's like digital cigarettes in this sense. We can't allow ourselves to be paralyzed on this issues because of a theoretical slippery slope. Whether Discord is going about this the right way is open for debate, and whether legislation solves the porn problem without introducing surveillance risks is also a good discussion to have. But the porn as well as the fraud and exploitation which always seem to accompany that industry need to go. Libertarians would be wise not to conflate the endorsement of privacy with an endorsement of porn -- most people support the former to some degree, but when people come forward with enthusiastic support for the latter, more often than not their motivation is addiction or profit, not a crowd the defenders of privacy want to be lumped in with.

safety1st··on Discord will require a face scan or ID for full access next month
The problem with the slippery slope argument is that it's a fallacy. That is the origin of the term, it describes a type of argument that's logically invalid. Yeah I am concerned that things could get worse and this might be the first step to broader censorship that we don't want, but a fallacious argument alone is unpersuasive to anyone who tries to form opinions rationally. Specific evidence needs to exist for the claim for it to be convincing.

Since slippery slopes are invalid by nature they're a type of argument that can be made for pretty much anything. If the case here is that a slippery slope is being used to defend pornographers and the "right to goon," I'm not on board. I think we have a long way to go to roll back porn's grip on teens and adults alike and reduce the harm it does to relationships, and this is just the beginning. Take for instance how Instagram at this point is basically a lead generation service for fraudulent OnlyFans businesses that sell parasocial relationships with a porn model's image where the customers aren't actually talking to her, they're talking to a team of guys in a basement in Eastern Europe somewhere. I think you shut down OnlyFans, you prosecute Meta, and to the extent where Discord is doing the same thing IG is, you prosecute Discord too. There's a long long list of things that needs to happen and shutting down the porn pipeline for teens on Discord is just the beginning.

safety1st··on Discord will require a face scan or ID for full access next month
I'm sure I'm in the minority here, but I read the announcement and other than the risk of a slippery slope into more invasive ID demands, I'm not sure I have a huge problem with it.

The default experience will be the "teen" experience - they list what that entails - stuff that's flagged as adult/NSFW/etc. is blurred out until your age is verified, which for most(?) people will require ID or face scan. DMs/friend requests from people you don't know take some extra clicks to view. Fine.

It depends on how broad the definition of adult content ends up being I guess, but I'm simply not convinced that requiring ID to view "adult" content is the end of the world. If that means porn, I'm 100% OK with it, put porn behind gates. It has become far too easy to access. It's 2026 and we now have a generation of gooning addicts out there who never have actual sex and it's basically a guarantee that they won't find partners or start families any time soon, exacerbating an already problematic decline in the birth rate. This is not a version of society or anyone's "rights" that I care to defend. You want to goon, show ID. That's how it was before the Internet anyway.

On the other hand if it means any speech that the platform deems to be "controversial" will be blurred out then my response will not be to submit ID, I'll simply limit how I use the platform. Anonymous speech continues to matter and needs protection. But Discord was never the entity that was going to provide that protection.

I mean Discord is a gaming chat room. Expectations should be set by that fact. I don't need a gaming chat room to be NSFW, or even host i.e. political speech really. I get that people have used it for more than gaming, but it was always pretty clear what it was. If people don't like that this gaming chat room no longer supports other uses, they should switch to an alternative.

safety1st··on AI fatigue is real and nobody talks about it
I'm not at all convinced that "break your concentration and go check on an agent once every several minutes" is a productivity increaser. We already know that compulsively checking your inbox while you try to code makes your output worse. Both kill your focus and that focus isn't optional when you're doing cognitively taxing work--you know, the stuff an AI can't do. So at the moment it's like we're lobotomizing ourselves in order to babysit a robot that's dumber than we are.

That said I don't dispute the value of agents but I haven't really figured out what the right workflow is. I think the AI either needs to be really fast if it's going to help me with my main task, so that it doesn't mess up my state of flow/concentration, or it needs to be something I set and forget for long periods of time. For the latter maybe the "AIs submitting PRs" approach will ultimately be the right way to go but I have yet to come across an agent whose output doesn't require quite a lot of planning, back and forth, and code review. I'm still thinking in the long run the main enduring value may be that these LLMs are a "conversational UI" to something, not that they're going to be like little mini-employees.

safety1st··on Microsoft's Copilot chatbot is running into problems
The comparison to the dotcom bubble isn't without merit. As a technology in terms of its applications though I think the best one to compare the LLM with is the mouse. It was absolutely a revolution in terms of how we interact with computers. You could do many tasks much faster with a GUI. Nearly all software was redesigned around it. The story around a "conversational interface" enabled by an LLM is similar. You can literally see the agent go off and run 10 grep commands or whatever in seconds, that you would have had to look up.

The mouse didn't become some huge profit center and the economy didn't realign around mouse manufacturers. People sure made a lot of money off it indirectly though. The profits accrued from sales of software that supported it well and delivered productivity improvements. Some of the companies who wrote that software also manufactured mice, some didn't.

I think it'll be the same now. It's far from clear that developing and hosting LLMs will be a great business. They'll transform computing anyway. The actual profits will accrue to whoever delivers software which integrates them in a way that delivers more productivity. On some level I feel like it's already happening, Gemini's well integrated into Google Drive, changes how I use it, and saves me time. ChatGPT is just a thing off on the side that I chat randomly with about my hangover. Github Copilot claims it's going to deliver productivity and sometimes kinda does but man it often sucks. Easy to infer from this info who my money will end up going to in the long run.

On diversification, I think anyone who's not a professional investor should steer away from picking individual stocks and already be diversified... I wouldn't advise anyone to get out of the market or to try and time the market. But a correction will come eventually and being invested in very broad index funds smooths out these bumps. To those of us who invest in the whole market, it's notable that a few big AI/tech companies have become a far larger share of the indices than they used to be, and a fairly sure bet that one day, they won't be anymore.

safety1st··on Two kinds of AI users are emerging
Well the second group in your taxonomy are very unserious, I mean that's fine, it's OK to use an AI tool for vibing and self-amusement, there will be an entire multi-billion dollar entertainment industry which will grow up around that. In my personal experience, decisionmakers who fell into this camp and were frothing at the mouth about making serious business decisions this way are already starting to get a reality check.

From my perspective the distinction is more on the supply side and we have two generations of AI tools. The first generation was simply talking to a chatbot in a web UI and it's still got its uses, you chat and build up a context with it, it's relying heavily on its training data, maybe it's reading one file.

The second generation leans into RAG and agentic capabilities (if you can glob and grep or otherwise run a search, congrats you have v1 of your RAG strategy). This is where Gemini actually scans all the docs in our Google Workspace and produces a proposal similar to ones we've written before. (Do we even need document templates anymore?) Or where you start a new programming project and Claude can write all the boilerplate, deploy and set up a barebones test suite within a couple of minutes. There's no doubt that these types of tools give us new capabilities and in some cases save a lot more time than just babbling into chatgpt.com.

I think this accounts for a lot of differences in terms of reported productivity by the sane users. I was way less enthusiastic about AI productivity gains before I discovered the "gen 2" applications.

safety1st··on Claude's new constitution
Yeah, this kind of gets to my main point. A prohibition against slavery very clearly protects the weak. The authorities don't get enslaved, the weak do. Who does a prohibition against "cyberweapons" protect? Well nobody really wants cyberweapons to proliferate, true, but the main type of actor with this concern is a state. This "constitution" is written from the perspective of protecting states, not people, and whether intentional or not, I think it'll turn out to be a tool for injustice because of that.

I was really disappointed with the rebuttals to what I wrote as well - like "the UNDHR is invalid because it's too politicized," or "your desire to protect human rights like freedom of expression, private property rights, or not being enslaved isn't as universal as you think." Wow, whoever these guys are who think this have fallen a long way down the nihilist rabbit hole, and should not be allowed anywhere near AI governance.

safety1st··on The tech monoculture is finally breaking
The root problem isn't really multi-purpose tech. It's the perennial coercive tendencies of monopolies being multiplied by their modern capability to update software in the blink of an eye.

If a company develops a monopoly in virtually any part of your life these days, and if a $1 network connected SoC can be added to their product, they can start abusing their position within a matter of months. The standard playbook is some combination of advertisements, notifications, and subscription charges (sometimes for stuff that used to be free!). None of those things are met with enthusiasm from consumers. But if the consumer has no other choice, it's almost a guarantee that the business will add them eventually.

Lock in and abuse. This isn't a new business model, we've just watched it spread from being a Microsoft PC thing in 1990s IT departments to pretty much everywhere now. (Speaking broadly about MSFT's business strategy back then, but they were also literally the first ones to try and shove unwanted Internet ads down your throat by streaming Active Desktop Channels on top of your wallpaper in 1997...!)

safety1st··on Claude's new constitution
I felt that section was pretty concerning, not for what it includes, but for what it fails to include. As a related concern, my expectation was that this "constitution" would bear some resemblance to other seminal works that declare rights and protections, it seems like it isn't influenced by any of those.

So for example we might look at the Universal Declaration of Human Rights. They really went for the big stuff with that one. Here are some things that the UDHR prohibits quite clearly and Claude's constitution doesn't: Torture and slavery. Neither one is ruled out in this constitution. Slavery is not mentioned once in this document. It says that torture is a tricky topic!

Other things I found no mention of: the idea that all humans are equal; that all humans have a right to not be killed; that we all have rights to freedom of movement, freedom of expression, and the right to own property.

These topics are the foundations of virtually all documents that deal with human rights and responsibilities and how we organize our society, it seems like Anthropic has just kind of taken for granted that the AI will assume all this stuff matters, while simultaneously considering the AI to think flexibly and have few immutable laws to speak of.

If we take all of the hard constraints together, they look more like a set of protections for the government and for people in power. Don't help someone build a weapon. Don't help someone damage infrastructure. Don't make any CSAM, etc. Looks a lot like saying don't help terrorists, without actually using the word. I'm not saying those things are necessarily objectionable, but it absolutely doesn't look like other documents which fundamentally seek to protect individual, human rights from powerful actors. If you told me it was written by the State Department, DoJ or the White House, I would believe you.

safety1st··on The recurring dream of replacing developers
I think you missed my point.

If you "quiet quit" you're still working for someone you hate. They still own you.

Instead, you could ACTUAL QUIT and start a business. Then you work for yourself, who you hopefully don't hate as much, and you have the power to define how things are done. So if you think developers shouldn't be replaced by LLMs or whatever, then you can... not do that. I have zero doubt in my mind that that will be a niche somewhere in the global economy for many years to come.

Also you might make a lot of money. Make enough and you're basically free from all these assholes. It doesn't actually take a ton of money when you own 100%. You pretty quickly get to a point where you can start telling off any wanker you want to, but you don't even feel the need, you are your own boss, so you just walk away from them and deal with whoever you want to deal with instead.

Trust me, speaking from experience, this is 1,000,000% better than "quiet quitting" which is pretty much remaining a corporate serf, just the most useless and pussy one in somebody else's room. Of course it's harder, most people who try it fail, real economics, real laws and real people will call you out pretty fast if you don't do things that people other than you think are valuable!

safety1st··on Gas Town Decoded
If gas town can actually do stuff well at any price it'll have a radical impact on how society is organized, because there are people out there who have practically unlimited money (billions of dollars of their own to spend, plus they can get the government to print more dollars for them if necessary; you probably already know who a few of these people are).

I've only started using coding agents recently and I think they go a long way to explain why different people get different mileage from "AI." My experience with Opencode using its default model, vs. Github Copilot using its default model, is night and day. One is amazing, the other is pretty crappy. That's a product of both the software/interface and the model itself I'd suspect.

Where I think this goes in the medium term is we will absolutely spin up our own teams of agents, probably not conforming to the silly anthropomorphized "town" model with mayors and polecats and so on, but they'll be specialized to particular purposes and respond to specific events within a software architecture or a project or even a business model. Currently the sky's the limit in my mind for all the possible applications of this, and a lot of it can be done with existing and fairly cheap models too, so the bottleneck is, surprise surprise... developer time! The industry won't disappear but it will increasingly revolve around orchestrating these teams of models, and software will continue to eat the world.

safety1st··on Scott Adams has died
Yeah I have come across it too, I have also met examples like the woman you describe. But we don't really have to rely on personal anecdotes. The rise of anxiety in young people over the last 20 years is well documented. Someone who's really determined to pick holes in this will say that doesn't prove causality, it could be multivariate or it could be other things completely, and they're right, we're probably not going to find a gold standard scientific study proving my point. But if someone thinks this increase in anxiety is not tied to how people react to speech, online and off, or if they try to handwave it away as unconnected to the broader social change I'm describing, they're being obstinate or they're trying to protect their sacred cows... for another example we have many many people of all political leanings (including apolitical) these days talking about how they've disappeared from public social media and retreated into private chat groups because the public discourse is just too dangerous. That is cancel culture. It is real. It has had precisely the deleterious effect on society which I described.
← PreviousPage 3 of 34Next →