Hacker News Reader

Comment by jagermo | Hacker News Reader

I get the appeal of agents, I do. This is the future stuff we always wanted. But I cannot get myself to give one of these things access to my bank account or allow it to do price comparsion and shopping without oversight. Or access to my email or chat history.

I just do not trust any of them, not with my money or with access to my conversations.

View on HN

What AI boosters don't realize is that we're not in the optimistic days of the 1990s anymore. We've seen that the prevailing tech business model is to offer convenience as a lure to lock in dependent users and extract value from them.

Unless these AI assistants are running on self-managed platforms independent of their developers, all I see is immense counterparty risk.

Absolutely. If you need any signal to know where Big Tech's head is at, just look at how quickly they became arms dealers.

Big Tech's head is at "being evil is just being competent", "democracy is incompatible with freedom" (for me to do what I want). It is also at "humans are obsolete" point. And while Zuckenberg specifically does not have cool quotes, he was at the "who cares about genocide" and "let show weight loss ads to teenagers who we determined have low self-esteem" points in the past.

Notes: actual quote is "If you're evil, you're at least competent. And if you're evil, you're not bad. And therefore, maybe you're actually kind of good because you're at least getting something done"

Quote was from Thiel FYI. https://youtu.be/B7yl7fEHeKM?t=840

Actual Demon Thiel

Impressive how every single statement is absolutely wrong.

You mean wrong because you don't believe this is how they are acting (likely because you are financially secure enough not to notice - part of the problem) or what they believe is dangerous and malevolent or because it's also nonsensical, despite being dangerous and worthy of reproach?

No, more like:

> If you're evil, you're at least competent.

Incompetent evil exists. This statement is wrong, B does not follow A.

> And if you're evil, you're not bad.

They aren't mutually exclusive, so this is incorrect.

> And therefore, maybe you're actually kind of good because you're at least getting something done

This presupposes that "getting something done" is a positive end in itself, which is not a given.

Each phrase is wrong. It's impressive, actually.

i think it was a bad analogy. he was talking about competent evil for eg ai enhanced weapons systems vs incompetent evil for eg middle easterners stoning people coz of religion.

both are evil but you can rationalise competent evil coz you're getting something real done like preventing the upending of the moral world order.

You're getting something real done like destroying moral world order, turning the western world into feudal techno fascist system. Which is his actual goal he stated multiple times.

Considering weirdness of Thiel believes, his praise and support for actual atrocities, including actually in middle east, trying contrast him with violent religious fanatics is weird. They dont contrast, they have notable similarities.

Weird that Thiel seems to think the “Antichrist” is a bad thing, given that quote. I suppose he sees it as unwelcome competition.

> What AI boosters don't realize is that we're not in the optimistic days of the 1990s anymore.

Comments like this are in a different reality than most consumers. Most consumers don’t care about things like avoiding lock in to a platform. If the platform solves their problem then they don’t have any reason to leave it anyway. They’re not worried if their data is used to show them more targeted ads.

Topics like this show a sharp divergence between what you read on Hacker News and how actual users operate. We already knew that people who don’t trust Facebook aren’t going to suddenly start using Muse. They were never going to consider it. For the people who do actually use Meta products, which is a customer base counted in the billions, many will not have any problem using these tools.

I see a lot of validity in your reply, but I think there is still nuance in how these priorities are expressed. How would you compare this issue to the Flock backlash, for example?

I think most people draw a strong distinction between privacy violations that are done "to them" and privacy issues with things they've chosen to use, even if they're quite similar from a technological perspective. I'd bet a lot of anti-Flock people have location sharing enabled on their phone.

And there is a very American anti-government aspect to it as well.

This is only true as long as they don’t think about the future or do risk management.

I asked my wife to not share any info with Muse and she said “Well they know everything already”. Well I was able to convince her not to with some stories about hallucinations of agents getting things horribly wrong as agents in the past, because the privacy concerns just didn’t phase her. Shes even more aware than most people since I’m very privacy aware. 100% it’s very echo chambery to claim that it’s not 1990 anymore. It’s like a historian saying that history won’t repeat itself because we all know what happened last time

My wife and I do not even allow Meta Apple and Google products or proprietary software in our home (unless owned by guests) because they are all predatory and endless alternatives exist.

Do you allow guests to use their Meta Apple and Google products in your home?

Begrudgingly if for short periods of time, on a dedicated wifi network, but if they are visiting my home I encourage them to interact with me and others instead of doom scrolling.

Try as I might some are so addicted and gone now that they just come over, doom scroll for a few hours, and leave. Maybe they feel like they get something out of just being in the same room as other people. It confuses me, but I will not force my views on guests.

What software / hardware do you use? How do you get around binary firmware blobs?

I use Linux and maintain my own Linux distribution. All software I directly touch and rely on is Open Source and does not spy on me. I do not own or use a smartphone. I do however permit the use of proprietary software for purely entertainment use cases, such as video games, but never for a -tool- that I use to do my job or live my life.

I even have a Talos workstation. Open hardware motherboard and open spec CPU both made in the US with fully open hardware. Also a Precursor which is similarly open.

I support these efforts as much as I financially can, but I admit open firmware/hardware is not a gap I have fully closed day to day. Yet. Deep down in some hardware is firmware I do not control, such as on the inside of my AMD GPUs. But I wallet-vote for those over Nvidia who require proprietary code in user space where it requires rights within my operating system.

> I do however permit the use of proprietary software for purely entertainment use cases, such as video games

How do you isolate those from affecting the rest of your machine? Do you use a different machine for entertainment? A different account? Something else?

I’m not judging or trying to catch you in some kind of gotcha, I’m interested in your setup in case the idea is something I could use myself.

All machines that can do things like play games are untrusted to the point I never even lock them. They are treated as game console.

I sometimes consider going scorched earth, but then I realize doing so will leave a conspicuous void which may also limit my options in the future.

Hmm. The areas where I ran into a brick wall on this were firmware for SSDs and hard drives. Doesn’t exist in open source form.

True but if it is immutable it still meets the FSF definition of an appliance regarded as a closed hardware problem vs a closed software problem.

Software Freedom means no one has more control of the hardware in front of you than you do. If no one can change the SSD firmware, then it is part of the hardware.

Now for firmware that can be changed, that is a different story and maybe the OpenSSD project gets us there. We will see.

Most consumers absolutely care, why do you think the biggest bipartisan issue is a massive national backlash against tech companies and a clear majority of workers being against LLM tools in the workplace?

These types of comments just show what a massive bubble you're in.

That backlash is based on conspiracy theories about data centers using up all the water and the legitimate fear of AI taking their jobs, not anything to do with privacy.

No the backlash is based on very real decline of material concerns, please try to speak in reality and talk to other humans outside of VC.

That's been happening for the past almost exactly 60 years, AI had nothing to do with the majority of it.

> a clear majority of workers being against LLM tools in the workplace? > These types of comments just show what a massive bubble you're in.

The fact you think it's a "clear majority" and not a "sizable, but still loud minority" of workers shows that you're in a bubble.

The vast, vast, majority of people are largely apathetic to AI as a technology at large, with general techbro-sentiment trending very negative (for all technology, across the board), especially towards SV/AI figureheads.

Well they will start caring when their teen with image issues starts getting targetted beauty ads[1], or when they get pregnant, want to abort, and get directed to an anti-abortion center instead[2], and afterwards they get served "abortion reversal pills" ads[3].

Or when they are hit by any other privacy fuckup like that.

[1] https://futurism.com/facebook-beauty-targeted-ads [2] https://www.bbc.com/news/health-61320202 [3] https://tech.yahoo.com/general/articles/facebook-made-money-...

Imagine we were talking about soda (or pop, or coke, or soft drink, depending on your dialect). Personally, I find it absolute insanity to be regularly consuming so much dissolved sugar especially during our ongoing obesity and digestive health epidemic.

But, “most consumers” don’t care about that. The beverage industry is insanely profitable, especially with recent forays into “energy drinks”.

Do people who drink so much dissolved sugar live without consequences? Absolutely not. They have a significantly lower quality of life and die much sooner than people who do not drink dissolved sugar.

I don’t see your comment as a valid dismissal. Just one more way that individual outcomes in our society are increasingly K-shaped.

Another way to say this is that your parent may be suffering from lack of imagination. Stuff like this can (and should) be regulated. Consumer protection is a thing most jurisdictions do. In fact there are millions of possible markets which consumers would absolutely fall victims to, but government regulators prohibit it because of the harm it causes consumers. Government regulators can (and should) apply the same logic to AI, but they don‘t. There is not really a good reason why they don‘t other then some ideological movement towards ever greater neo-liberalism, deregulation, and laissez faire capitalism.

Expanding on this thought (too late for an edit), there's a nice simile to be had:

  - technologists : software that is bad for privacy
  - nutritionists : food and drink that gives you diabetes

So, I'm no nutritionist, but I pay attention enough to pay attention when they say I shouldn't drink a liter of Coke every day. As technologists, we should carry the torch and speak loud-and-wide about the dangers of software that is bad for privacy.

> I don’t see your comment as a valid dismissal. Just one more way that individual outcomes in our society are increasingly K-shaped.

I was dismissing the claim that the product will fail because the world doesn’t like it.

You changed the subject to something else.

You’re right, I was responding more to the sentiment of the overall thread than your comment in particular. Mea culpa.

Most consumers do not want or need agents which also shows how much of a bubble tech people are living in.

How do you explain the brand segmentation Facebook is doing with Muse? They're keeping the Meta / Facebook brand very far away from the Muse product.

Doesn't that imply Meta itself believes users care?

If we’re talking about the average consumer, they’re increasingly hostile towards ML in general and Facebook in particular. The majority want the tech regulated stat, so you might want to think twice before appealing to the masses.

Consumers do care about these things, a lot. The issue is they have a lack of choice and transparency. Everyone is basically just OK with tech being shit because all tech is shit, so now you're just choosing shit.

Take smartphones. Don't want platform lock-in? Go fuck yourself. You're options are Apple and Google, who both maximize platform lock-in. You could always go live in the woods, that's freedom of choice baby. Welcome to the free market.

Of course, this is something you'll have to decide after reading the hundreds of pages of terms of use, privacy policies, and EULAs to learn how your data is actually being used. Oh, and they all update weekly. And you automatically consent.

The truth is that nobody trusts Facebook. It's a bargain. If you want to do certain things, you're forced to relinquish control. It doesn't matter if you trust them or not.

Me and 4 Billion happy users don't care.

HN must understand that they are not a representative sample of anything.

Most of the planet refused to wash their hands or use basic sanitation until about 150 years ago even though they had clear reproducible evidence doing so saved countless lives 50 years earlier. Most just considered doing the bare minimum to save lives as too much work. Also we used to advertise cigarettes and alcohol for consumption by babies in the 50s.

It is reasonable to assume the majority of humans are intellectually lazy to the point of killing themselves and their families if told to with the right marketing.

We should all make our own individual well researched choices and not ever buy into the worthless argument of "everyone does it".

False analogies won't make your case.

There are plenty of happy, successful people who didn't die because they used a Meta product.

In fact I'd even argue that all things equal, a Meta user is happier than a paranoid-meta-hater who worries every minute that someone is going to serve them an Ad that fits them

"oh, the horror!!!. How am I ever going to recover from the trauma and financial ruin of seeing a targeted ad."

>successful people who didn't die

Did you just move the goalposts to “it didn’t kill me”?

We’re talking about society-wide effects here, including harm to children for which Facebook is paying 17 billion dollars.

https://oag.dc.gov/release/attorney-general-schwalb-announce...

And those children are $17B richer today.

The ones that did not commit suicide get a tiny tiny fraction of that, yes.

Your comment implies no harm matters so long as people can sue for it later?

For a supposedly science-based cohort we aren't using any science aren't we?

There is ZERO proven causality between social media usage (specifically Meta products) and long-term happiness of a person. All the social science are spurious correlations that can be attributed to anything (including higher education where hysteria about the modern world reins)

As far as $17B settlement, it has nothing to do with causality. It's just some $$$ set aside for Meta to remove operational headaches.

I could even say Meta-haters make more irrational life decisions compared to normal people who just browse Instagram and use WhatsApp without paranoia

Did you just accuse others of not being science-based, while casually handwaving studies away?

Can you point me to causation studies please? or are you the type who believe in social sciences (using self-reported, biased data) with 0.10 R2 values and think that's the smoking gun?

Nope, it’s on you to either provide counter-studies first, or even just provide one solid, concrete example of what issue you have with these studies, rather than just handwaving them away as if such a rhetorical flourish can substantiate any sort of argument.

> There are plenty of happy, successful people who didn't die because they used a Meta product.

And plenty of teens that never became happy and successful people because they used a Meta product.

These companies are predatory, and have nothing to offer us we cannot easily setup at home these days with privacy and sovereignty.

[citation needed]

Note: Unless you spin an alternative universe there is no way you can prove causality of social media usage => happiness/depression.

Meta would not have paid out if they saw the evidence and thought they had a chance in hell at winning.

But it is the undeniable opinion of most mental healthcare professionals that giving teens an endless stream of content about unattainable body images or suicide content... results in more suicides.

Meta -knew- this and then fired the team that told them about it and doubled down on pressing the gas. They knowingly and substantially increased the chances of teen suicides for money.

And people want to give these monsters -MORE- data and power? I cannot comprehend anyone defending this company still.

Meta had to pay out because the entire country is brainwashed with anti-Bigtech hate and 99.99% are clueless about correlation != causation and there is no point in fighting the court where the media will make sure to amplify 1 in 100 Million cases as thought it's the norm.

I mean some of the smartest people on HN fall for it, where is the hope for general masses?

I am at least encouraged to see even one of the last anti-privacy chads left is admitting the surveillance capitalism corpos are losing the war of public opinion.

Every time their stock price takes a hit my hope in humanity gets a little boost.

All proprietary technology will fail. I really wish it would fail faster though.

> "oh, the horror!!!. How am I ever going to recover from the trauma and financial ruin of seeing a targeted ad."

Well yeah that... and the spread of misinformation on Meta platforms leading to deaths from COVID. And the genocide in Myanmar they caused.

Information and outcomes are not separate. If you think the words you see, which are very carefully curated, truly have no effect on outcomes then you must have never attended a history class. Like, ever.

Not at all, if you are happy, go for it. I might look into something like hermes, that seems to fit my bill more.

I read "the tesla files", a book about a whistleblower leaking a ton of internal tesla files.

They were discussing how tesla manages problems with full-self-driving.

It was interesting how this realm of problems was viewed.

If there was an accident while the car was driving itself due to some glitch, the police at the scene put all the blame on the driver.

The book had a different viewpoint. obviously there was a tesla bug in full self driving, and they kept their mouth shut.

meanwhile society/government doesn't even think of this and puts responsibility/blame on the user.

AI seems to be an amplifier for other problems that we never fixed, more than an authentic source of problems on its own. In this case it's amplifying that we never really trusted the cloud to begin with.

We have a lot of deferred problems to go back and solve before all these dreams can come true.

I think WallE is the perfect reference for the world AI/oligarchy are running towards. Not whether it exists, but that' dependency+environmental destitution.

I agree with this, but I’m wondering where these self managed platforms are for the average person who doesn’t want to spend hours fixing OpenClaw upgrades. Surely there has to be an easy to use Muse/Dots/GrokBot self hosted solution for the privacy conscious?

I see the value of the tool but I would never use it from Meta. Would need to be self hosted with a very structured framework and guardrails so that even my local model couldn't accidentally wire 50k to a Nigerian prince or whatever the latest email scam is.

I see everybody freaking out about unfettered payment access to one's bank accounts and I keep wondering why people don't jump to the obvious solution of simply only giving it a single purpose or limited fund credit card number from privacy.com or something.

Am I missing something with the concern about ability to constrain the blast radius?

I would call that kind of thing "a very structured framework and guardrails". You'd have to come up with entirely different solutions if you want the agent to receive payments or monitor your budget.

Man, I wonder how many facebook marketplace sales will be something like "In order to make sure you're a real person, before we can meet up so I can hand over this brand new macbook for $500 I need you to send me $10 electronically. It's really just to verify your commitment". With the right prompting, I'd be entirely unsurprised if Muse will just send over the $10, and of course the seller ghosts the buyer -_-

I've sold a few things on marketplace and have got some strange requests from buyers that i think are AI. I was selling a piece of exercise equipment that has a screen attached to it (rowing machine) and got a request to take a picture with a tape measure to verify the screen dimensions. The product is only sold with one screen size and it was plainly visible in the listing pictures too.

That could be either lazy buyer or someone who's had a seller tell them wrong dimensions or wrong model for the item. Someone I know from college told me a couple years about how they lied to a buyer about the cheap wayfair furniture was actually part of a sample sale from a big name brand. I usually ask the seller for any info thats make or break for me, but I also reverse google image search the listing photos to make it easier for myself though.

Classic GPTism that. It often asks me to take pictures of things with a tape measure across it.

Been using it to help with a refurb of an old cottage as kind of a super google for old fittings etc.

And I see this few becoming more common. May I ask what agent you see? Most likely good for self-hosting?

I’m in the same boat. After witnessing context rot and inference collapse, I do not trust any LLM with mission critical work. Not Jev, not grok, not fable, not Opus.

What extent does that happen for you? I have got very good results with self hosted qwen3.8 flash next at q4 and even with qwen3.635b on a laptop. I don't keep it running forever on every single repo, its ok to leave notes in agents.md and let it search that instead of the entire history staying in context.

I’m not talking coding agents. More so agent harnesses and using LLMs for decision making

heh I got a completely stupid error from gemini about some apache configs, when I pointed it out, the llm apologised, said the line(s?) should look like this, then posted the same two lines uneditted haha

The Qwen models, especially when quantized, can be really bad about just trying things and seeing what works. If you’re not watching all the tool calls you may not see it, but it’s kind of scary to watch them just bump into wrong decisions and backtrack.

They also have a bad habit of accidentally building URLs that hit Alibaba infrastructure, likely because their training environment had them use those URLs. If you haven’t watched the outgoing network requests you might be very surprised at what your Qwen agents do sometimes.

> [Humans] can be really bad about just trying things and seeing what works. If you’re not watching all the tool calls you may not see it, but it’s kind of scary to watch them just bump into wrong decisions and backtrack.

>This is the future stuff we always wanted.

Speak for yourself. I hated the idea of agents ever since I first heard of it back in the 90s or 00s. For me, the most valuable feature of computers is predictability. I want tools, not agents. A computer should augment my own skills, not replace them.

I mean, fair. I want something like the TNG Enterprise computer, that works for me within the parameters I set. That, for me, is cool.

Pretend you have insifficient skill to satisfy your desires, even after augmentation.

You are now dead center of the "AI" agent's target market.

The missing piece for connecting an agent to your credit card or other financials is financial liability.

If Meta/OpenAI/etc would guarantee they’d compensate you in full for anything that wasn’t supposed to happen —and had an established track record of doing so— it would make trusting agents to make financial choices easier

“ If Meta/OpenAI/etc would guarantee they’d compensate you in full for anything that wasn’t supposed to happen —and had an established track record of doing so— it would make trusting agents to make financial choices easier”

lol… talk about delusion.

First you would need to establish some sort of interaction. Probably with the same AI that spaffed your cash.

I think the solution is instead to require purchases to hit a hard harness restriction, so you must click yes for money to be spent.

Assuming financial liability is an option, but because the risk profile of having an agent making purchases for a cardholder will never be zero, the companies making these guarantees will simply price the risk in to the cost of service. It’s the same basic mechanic at work with credit card reward points, the brands are all fighting each other for a share of customer debt so they slowly increase the rewards rates to one-up each other, and make up the cost on the back end by increasing merchant fees. It’s a dumb, unnecessary price war that ultimately increases the cost to the consumer and drives up the profits of the businesses providing risk/cost-burdened services.

Sad to see such paranoia, agents can genuinely be an agent of good and positive change.

Last night I had ChatGPT go through an old email account and surface memories that I literally forgot. People who I have fond memories of. Yea I could have spent time querying gmail and digging but the agent did it in a way that really resonated. I don't care if Sama has my emails now. I do care about connecting with my past and enjoying the memories of a time long past.

Paranoid scolds of HN want the average person to be deprived of value like this for some reason.

I affirmatively want that exact scenario not to happen, yes. I'm opposed to the automation of nostalgia, I don't think it's good nor positive.

> I don't care if Sama has my emails now.

I am perfectly prepared to believe you had a nice interaction with the robot. But this is an insane thing to say!

My email is gigabytes of garbage and a few useful bits of information that are useful and have personal relevance to me. You personally, dear reader, could probably read all my emails and I wouldn’t give a shit. I just checked my sent mails for the last few years, it’s like 90% sending absence excuses to my kid’s school. Google already has it all, so I guess I could just wait for them to get off their ass and make a useful equivalent to Grok Bot and friends.

I think that's fair and not a problem if you feel comfortable sharing that. But I think what op wants to say is that not sharing your email should be the default state, and only giving that when you want to get value in return like in this case with your agent

Independently of the actual state of the art, and the social implications and everything, my main takeaway from Her was always the initial setup of Samantha (huh, now that's some eerie nominative coincidence), wherein the AI curates his email inbox. That is the mundane chore that's always been my dream AI use case.

https://www.youtube.com/watch?v=f9Hg1x-Ctlw

I made an agent to derive and iterate on a stock trading strategy for me. What I did was give it a virtual bank account that models your typical brokerage account. So my agent reads the news, financials, markets, all the information it can gather to decide what stocks to buy/sell and manage its portfolio. Then what I do is just follow along with real money.

I could see shopping being the same way, let the agent shop for you or identify good deals on things you want but then push them to you to make the actual purchase. Like you, I'm certainly not going to let an agent make a real purchase for me unattended and I doubt any merchant is going to have much sympathy for "my AI bought this by mistake".

It's evident that agents still can't be trusted to transact on your behalf. Too many things can go wrong.

I trust mine considerable, but only because I can run it mostly offline on hardware I own that lives in my garage. It baffles me that I am the only technical person I know that exclusively uses AI this way.

I am a technical person and while I appreciate the challenge of buying the hardware, setting up the software stack and managing the system, I can't figure out any use cases that would justify the work. This is the main issue I have with Muse, Dots, etc. There just must be something about my routines that don't align with what these companies are expecting their users to do. I do take advantage of the investor subsidized coding agents, but it will take years of my usage of their services to come close to initial start up cost of my own system.

I am very curious as to what other people of using these agents for? Some of the use cases I hear; comparative shopping, travel planning, etc. don't appeal to me. Our people using them to manage family life issues like school schedules, meetings, etc? I feel like I am missing out, but I also am not seeing the greater appeal yet.

Once you have a local agent you can actually trust is not sharing your conversations anywhere, you just use it like a really smart search engine, and your need to use a web browser mostly goes away.

Some of my recent use cases...

Help me find cheap flights and a hotel and plan some interesting sights to see because I am speaking at x conference on x date.

Download all my favorite music from x streaming service as FLAC files locally, then convert them to OGG and put them on my portable music player.

Download all my youtube subscriptions locally to this folder and strip out any ads or sponsored content.

Here are a list of all the books I own. Can you find me DRM-free epub copies and put them on this e-reader for me?

Here is a picture of an object next to a measuring tape. Can you 3d print me a holder perfectly sized for it based on a openscad file i can easily tweak?

I am really stressed right now. Can you help me talk through everything I have to do and help me rationally prioritize? Can you remind me and keep me on track and be my fake boss for today?

Can you look through my email and keep track of any appointments or invites so I stop missing meetings?

Thanks for the useful tasks. If I actually had your setup, I might do those things. Like you, I am not going to ask Meta, et al. to do those things. I also am hesitant to invest in setting up a new system. However, I do have a M2 Mac Studio, so maybe I am closer to being able to run a local system than I thought.

The cool part when you have a local decent model is you can ask it to write automations for you that you can review, and see what works for you. Can keep stacking new requirements.

I mostly use jcode in part because of the self-dev mode where it edits its own rust code and restarts itself into new binaries seamlessly to give itself new features.

I follow the same approach (finding LLMs useful but only running open models on my own hardware) but I've been limited by also not letting them into my email and being pretty restrictive about their internet access. Would you mind sharing what stack and agent harness you use, since your other comments make it clear you're highly restrictive about what kinds of tech you run at home and I assume you vetted it thoroughly, and it sounds like it's working well?

I'm wondering specifically about things like what kind of tool or restrictions you use on web browsing to avoid being flagged as a bot and how you guardrail access to your email (to avoid, say, the LLM deciding that "help me organize" includes telling people you want to reschedule). I use Qwen 3.8 26b at home and find it very useful for some tasks but when I think of using it as a general assistant rather than coding assistant, I keep thinking of times it has gone wildly wrong (e.g. it once spent 20 minutes trying to get around Supabase authentication because it couldn't figure why a certain call wasn't working)

My go to right now is jcode, and doing a lot of one-shot experiments. Like I have a session being my MPD DJ as my local spotify right now.

That said for persistent things like interfacing with homeassistant etc I am playing with zeroclaw.

I run the same model as you for coding. For more sane automation tasks i recommend doing everything possible with MPCs so you can tell it to only use existing tools. ymmv

By no means is my setup well oiled yet. Different vms and systems doing different things. As a QubesOS user jailing an agent for one job to one VM is easy, though I am working on an enclave native operating system that should make this easier for most people.

Just dropping in to say that this is no longer that big of a problem. There are hosted solutions only for agents from DO, Nous, etc But also more generic self-hosting solutions, where you can host a lot of stuff and also agents, like moose

I think as people realize that self-hosting is a good option. Good services will appear more

> allow it to do price comparsion and shopping without oversight

You can get it to do the price comparison part without having it do purchases.

I downloaded Muse, told it to find me some underwear to buy, asked some questions to narrow the options down, then once I came to a decision I went and purchased it myself.

Muse is basically OpenClaw, with all the pros and cons. Except I really don’t trust Meta to get any of this right.

Personally to me this feels like another classic case of starting with the technology and figuring out where to sell it as opposed to the customer experience.

It feels and seems too forced.

My wife has scheduled a visit with a physician and there was a younger person in there. They legit said “haha I don’t know what it is” and used ChatGPT for diagnosis.

I have no doubt they would have no problem outsourcing everything to agents.

That's some Idiocracy level bullshit.

Did they reach a good diagnosis by using tools to provide themselves additional information to process?

Or did they make a fatally wrong diagnosis that was only discovered because the old doctor whipped out their medical encyclopedias? You left us hanging without the relevant part of the story!

> Or did they make a fatally wrong diagnosis that was only discovered because the old doctor whipped out their medical encyclopedias?

You’re assuming you can get access to a qualified doctor in Europe, which isn’t always the case.

It was the same diagnosis my wife got from ChatGPT two weeks prior.

I think 90% of health issues are just hand-waved.

Don't worry, I'm sure eventually you'll simply be required to give one of these things access to your bank account and that decision will be taken out of your hands.

When that happens, I'm sure banks will lock out secure local models for "security reasons" like an unlocked phone.

Sorry, no. This is not the future I wanted. I do not run any agents nor ever will.

I find all this is just the next-gen privacy violation, disguised obviously as: "Oh, WOW an agent in the computer is doing all this for me". Bullshit.

It seems that never is enough with these ever thirsty companies, they keep pushing the limits and surprinsingly a lot of people do not care at all nor value the implications of having an arbitrary process running and doing pretty much whatever the agenda of their creators are.

This outcome is always likely/ predictable.

We are an opportunistic species. We don't have perfect knowledge nor thoughtfulness.

We need to protect each other.

We don't.

Now add Meta to the mix and stir.

> This is the future stuff we always wanted.

> (…)

> I just do not trust any of them

So… Actually not the future stuff we wanted? That’s what bothers me when people say “we have the Star Trek computer”¹. We clearly don’t, because if we did we could trust it with a high degree of certainty. Instead what we have is a computer we must distrust to a high degree. One of the two crucial variables is flipped.

¹ I’m not saying you are saying that, but several people have expressed that sentiment on HN.

You don't have to, there is a world of things you can do with them without giving them access to your email or bank account.

"Hermes, clone this android app to my computer, add this and this feature and fix this annoyance and then rebuild it and push it to my phone"

"Hermes summarize this YouTube podcast and push the summary and transcript to my Obsidian vault"

"Hermes check mg obsidian notes for when I last wrote a blog article about Neovim's integration with AI agents"

*Hermes, you have an API token that allows read only access to the Zabbix monitoring system, check that and then use the Proxmox API token that only allows you to do limited actions to reboot the VM that is having issues"

Etc, etc,

YouTube transcripts are increasingly if not completely blocked now. It used to be great to grab content from long tops. I did not have time to watch but wanted to learn about. Now. The only way to do it is to manually copy and paste the transcript information or use computer use. But I cannot easily do it on my VPS it seems unless I'm missing a trick.

You can transcribe in seconds on CPU for free. Check out yapsnap repo, its free and open source

I just have Hermes download the video using yt-dlp and transcribe it with a local whisper model using a local whisper server if the transcript is not available through the API, whisper base models works fairly well on CPU only servers nowadays.

Yeah, just waiting for the AI-enshittification when they give it to the growth hackers to drive VC returns at the expense of utility.

I just can’t even fathom it. LLMs have no judgment, and it’s still so easy to “trick” them. If you forced me to decide between granting the ability to conduct financial transactions on my behalf to an LLM agent, or to my family member with Lewy body dementia, I’d have to have a serious think about which one is less risky.

The right model for this usecase is a virtual card with a fixed budget, these exist.

I’m sure someone is going to lose their Robinhood account or savings but with due care this is avoidable.

And I think that's fair. On one side is their abilities, or lack thereof, but on the other side, it's the privacy implication with something like Muse.

I think for the former we can wait for bethel model, have better skills or give them small enough tasks. But for the latter, the only solution is honestly just hosting your own agent and all this having access to the data it creates, what it runs and what it reads.

Reply on news.ycombinator.com