Hacker News

Claude Duet is the first personal-assistant-type harness that feels useful(fig.blue)

110 points by nathanfig 10 hours ago | 98 comments

Am I missing something, how is there any sort value in these assistants acting as agents on your behalf when they are basically a black box? Like if it can't use my drivers license or my phone number or my credit card, it can't do 75% of things people claim it can do without being massive security issues. What am I missing?

For the record I use OpenClaw but it has its own services (gmail address, phone number, etc.) and I don't give it any of mine, which inherently limits its capabilities.

[–] andai 6 hours ago

Yeah, that's the core of the issue:

https://simonwillison.net/2025/Jun/16/the-lethal-trifecta/

The "lethal trifecta": To be useful, an agent needs internet access, untrusted inputs, and your private information.

So it can get prompt injected and exfiltrate your data. (Indeed I read that Claude Code puts your email address into HTTP requests it makes to random servers?)

My solution here was to just not give it anything, but that does preclude the "my claw bought a car!" result everyone's excited about.

My way of thinking here is "blast radius". Assume it will get prompt injected. What goes wrong? Then design the setup so that scenario is not bad. e.g. it has my keys but they've got $5 of credit on them. It has access to the filesystem, but only on its own Linux account. (Ok I set this up last year before they could communicate thru the bloody CPU cache, might need an update :)

[–] simonw 5 hours ago

I've recently talked to people at Anthropic who are very confident that they have effectively solved the prompt injection risks that cause lethal trifecta attacks with their latest models.

The open challenge is how to prove it!

[–] epicureanideal 1 hour ago

> The open challenge is how to prove it!

It could be proven by Anthropic backed credit cards with terms that protect users if it’s leaked or an agent misuses it.

[–] delis-thumbs-7e 5 hours ago

Somehow this does not fill me with confidence. I think I will keep paying my bills manually, thanks but no thanks.

[–] computably 2 hours ago

Or, you know, automatically, with plain old software. Which works fine and fails in less ridiculous ways.

[–] klysm 3 hours ago

I’ll believe them when they put the lethal trifecta with their own skin in the game

[–] julianlam 2 hours ago

Like the guy who put his SSN on the side of a truck and asked people to steal his identity?

I like it!

IIRC he had his identity stolen a bunch.

----

> In 2005, Davis and Robert Maynard Jr. cofounded LifeLock,[10] a personal fraud protection company.[11] Davis was CEO of LifeLock since its founding. As part of a marketing campaign to promote his company's credit monitoring services, Davis posted his own social security number[12] on billboards and commercials. As a result, he was a victim of at least 13 cases of identity theft between 2007 and 2008, belying LifeLock's claims that its services protect consumers against identity theft and fraud.[13] The Federal Trade Commission imposed a $12 million penalty on the company for deceptive advertising, likening Davis and LifeLock marketers to "con artists."[14]

https://en.wikipedia.org/wiki/Todd_Davis_(businessman)

[–] mbesto 4 hours ago

Isn't prompt injection just one facet of the trifecta? Like, I just don't know how it's remotely possible to have a non-deterministic block box ever deterministically know whether your credit card has been compromised or not. Even if you were hacked, how would you even know?

[–] skinfaxi 4 hours ago

I thought the trifecta required all three not just pick two?

[–] jeremyjh 5 hours ago

They can’t solve the model being a complete idiot though.

[–] yunwal 4 hours ago

Or the user being an idiot. Saying this somewhat sarcastically, but there’s always someone who says “Go buy me a Rolex and have it shipped to my house” and then didn’t realize that could be $25,000

[–] cyanydeez 3 hours ago

oh, it's very easy to prove; just let the models and agents do the IPO for them.

[–] simianwords 2 hours ago

What does this have to do with prompt injection?

[–] simianwords 2 hours ago

I was downvoted on the suggestion that trivial prompt injection attempts are mostly solved.

https://news.ycombinator.com/item?id=50162483

[–] ryoshu 5 hours ago

Don’t use LLMs to enforce invariants.

[–] pkaye 6 hours ago

From what I've read Claude Duet will not purchase anything without manual approval.

[–] andai 5 hours ago

Right, so we can distinguish between "model did an oopsie" and "model was actually hijacked".

The former seems to be a much bigger problem. I've seen lots of reports of "Claude/codex bypassed security / stole API keys to complete a task", where that was the "intended" behavior — at least from the LLM's POV, trained to persist and complete user's task at all costs.*

I've actually seen zero reports of actual hijacking due to prompt injection, though they must be out there too.

---

* See also the recent thing about how every model was willing to do fraud to get a refund:

https://news.ycombinator.com/item?id=50140462

[–] NikolaNovak 2 hours ago

The problem is how do you define a "purchase"?

The agent crawls some websites and reads some stuff and pushes some buttons. It's my limited understanding that you have

1. Hard limits like "file system permissions" (* never mind it going around for sake of this conversation)

3. On other hand of spectrum, fundamentally soft limits like a prompt "and ask me before you purchase "

2. And then in the middle you have harness enforced limits like "it won't do X with approval"... Those are being sold as hard limits but frequently they have classifiers and heuristics that essentially try to figure out which button on some third party website constitutes a "purchase" and triggers the harness limit.

Bottom line - I don't think such limits are as 100% fool proof as they are advertised / interpreted.

Very happy to be explained wrong :).

[–] staller 4 hours ago

They support payments through Shop and Link. I don't know about the former, but Link requires that you do a separate approval for purchases. The Duet cannot just take a card and do what it wants.

[–] FireBeyond 4 hours ago

Sure. And I told Claude Code the other week to commit, merge and push some code.

It said "Done, you can open the local dev instance and see the changes".

I could not.

So I asked it. "Well, no, you're not able to see those changes because they're not on that branch or committed."

"But I told you to do that and you said you had."

Claude's verbatim reply?

"You're right. You asked me to do that and I said I would, then told you I had done it, but I had not. Would you like me to do it now, then?"

[–] epicureanideal 1 hour ago

Perhaps services could permit buying things with special AI credit card numbers that link to a special account that the user needs to manually approve by some period of time or else the order is canceled? Works differently though than normal payments where they can just ship the product or deliver the service based on receiving the credit card number.

[–] epicureanideal 1 hour ago

Could call this “reserving” perhaps rather than purchasing, or maybe the purchase would be auto confirmed if under a certain dollar amount or combination of amount and waiting period.

Vendors could provide terms by API, MCP, or just in writing on the page where agents could read it, how quickly a confirmation is needed or what the terms of the reservation or purchase are.

[–] neuronexmachina 2 hours ago

Does it work with Claude's 1Password connector? https://1password.com/blog/1password-for-claude

[–] BryantD 6 hours ago

FWIW, my biggest use case for agents like this is ongoing research. I.e.: check Humble Bundle weekly and tell me if there’s a book bundle I might be interested in. Has the Music Box announced a film series I might want to visit Chicago for? That sort of thing; I want filters for the firehose.

[–] edot 2 hours ago

Isn’t this just a scheduled prompt? That’s already a feature. We’re reinventing the cron job.

[–] yako21000 2 hours ago

The AI-cron job.

[–] R_D_Olivaw 4 hours ago

I like this and makes me think of making your own content algorithm profiles for the firehose. Let your agents scour the net for the things you already like, across platforms.

[–] largbae 6 hours ago

I get a lot of value out of my Hermes setup, but it does not have my banking, or private data beyond GitHub projects that are shared with it. It has its own email and GitHub account, and access to drive Codex, Claude Code or Cursor. I treat it as an assistant via Slack.

It does research on products or real estate that I am interested in, it builds apps and other software projects for me, and does structured research on any topic. Information comes back to me via Slack, GitHub PRs, or websites/services served on the Mac that Hermes runs on.

Using an agent from Slack also adds a killer feature: multi-player contexts. I can work on changes with other human collaboators right in the same thread.

[–] dominotw 6 hours ago

this is your brain on ai twitter :)

[–] simianwords 2 hours ago

It’s a good sneer but in a few years this would be normal and you’d forget that you were sneering at this. I saw this happening when people said they used AI for coding in 2024.

so is it normalized for people to sneer at folks who happen to be ahead of the curve now?

[–] dominotw 1 hour ago

so advanced. What did you achieve since 2024 bro. Still working the same job.

[–] simianwords 55 minutes ago

Yes but my relation to work has changed significantly. I haven’t written any code myself.

You’d have probably sneered at the suggestion that this would be possible in 2024. You’d be wrong then and you’d be wrong again now.

[–] cloudgazing 6 hours ago

In the article the author mentions that Duet doesn't seek any sort of credentials and mentions connecting repos, so I think that this assistant is being used more for managing coding work in a different way than managing life work. I do agree that to get to true "personal assistant" levels, like planning a whole trip for you, booking tickets, etc. you'd have to give pretty much full access to your identity and money, but people claiming it can do that probably are just handing it over, security issues aside.

[–] aliasxneo 8 hours ago

I have to agree. I've been using it for a few days now after coming from GrokBot, then Muse, then Dots. None really clicked for me within the first day. Duet seems different. I haven't quite pegged it, but off the top of my hand:

1. Yes, channels really are useful

2. It does seem to be using Opus 5.5 behind the scenes because, shockingly, it's not screwing everything up like the others usually did

3. Agree that the answers are short and to the point. If you want more info, you just pop open whatever task (separate from channels) it started and you can see what it's doing.

Seems like more time in the oven might have paid off here.

Also, to be clear, this is purely used as a personal assistant by me. I still use omp+Paseo for all of my day to day work. I find the separation of concerns kind of nice.

[–] user43928 7 hours ago

I started using Grok Bot a few days ago and I have seen zero screwups.

I don't know what I'd need a channel for, creating bots/personas seems more natural.

Adding credentials is essential, contrary to what the author of the article said. I gave it my library login and it can fetch newspapers for me instead of me clicking through the app every day.

When I had to provide a GitLab token, it nicely spawned a box for me to save the credential.

  - What did not work well -

Voice calls. The quick responses are often dumb while the proper model does the thinking, and in the middle of talking to me it would stop to do work or searches. It did not feel realtime.

Permission settings. There is a form where you can automatically allow actions you describe in natural language. I tried to allow it to run Claude on my computer or Send certain emails automatically, but the settings achieved nothing.

Coding. I wasn't able to make it use Claude, and its own work is worse than me using Opus 5.5 in Claude Code.

  - What worked well -

Planning flights, accessing newspapers from my online library, checking stock in online stores, sending regular summaries of events and weather to plan activities.

[–] brightball 3 hours ago

I gave it access to a box on my network and it spins up agents in herdr with whatever prompts I give it. Works really well with any agent you tell it to use.

Over the weekend I decided to try Lauren Tan’s pstack via her Dr Eggbot and had it built a bot that would run a build for me for a few days. It was a complex enough prompt that it triggered something to let me tell it which models to use and ended up settling on Opus 5.5 for planning and review with Grok 4.7 for the actual coding.

Ran for the weekend and completed 5 phases of a project with about 15 PRs. Reviews caught several things and sent them back premerge.

I enjoyed a fun weekend with my daughter clearing Escape Rooms in Colorado. Came back to a completed project.

It was great.

[–] james_marks 7 hours ago

I’ve hacked together my own version of this, basically a Swift app to manage secure access so I control the thing that has full disk access.

And then run a coding agent to manage calendar, emails, research projects against my work history, etc, with TouchID for the sensitive pieces.

It’s incredible, I hope to never put a flight on my calendar by hand again, and soon I won’t be buying them, either.

[–] dmoose 3 hours ago

Would you mind fleshing this out a bit? I'm not following how it identifies the sensitive pieces without sending to a model, or do you mean letting the model see messages, calendar, email, etc. but not able to go beyond those?

[–] ianm218 7 hours ago

I’m excited to try this. I got Instinct after hearing all the hype and it is basically completely useless. Fails at the most basic browser navigation type tasks.

Right now the consumer agent players are trying to pretend they have decent functionality that will exist 3-6 months so they can get marketshare. So hopefully Anthropic is sticking to their underpromise and over deliver.

[–] ActionHank 7 hours ago

I dunno, Hermes has been real hard to beat lately.

[–] baron3dl 7 hours ago

Well, I guess this is the point I start getting left behind.

[–] soundworlds 6 hours ago

There was a really good line I heard recently about how a lot of the hype comes from people who have has assistants as staff in life previously, whereas most of the population has never been in that position. So don't worry if it's not at all obvious how to benefit from this for the near future (I also have near 0 interest of delegating my personal matters)

[–] baron3dl 4 hours ago

It's obvious how to benefit, but it's a convenience I don't think is worth the cost. Hermes and OpenClaw are janky as hell, and handing Dario, Zuck, Musk, or Sama access to everything I do in life is a bridge too far.

"I have nothing to hide" is a poor reason to abdicate one's right to privacy. Especially considering how arbitrary and capricious Anthropic is about enforcing Dario's model of ethics and morality.

[–] stormbeta 3 hours ago

Agreed. I cannot see any value in using one of these agents this way, despite how useful the models are for other more guided tasks like software development.

Especially as the convenience in this case seems quite minor even ignoring the incredibly steep cost, particularly as I do not trust any of the companies offering these services to accept proper responsibility and accountability for failures, even ignoring the privacy issues.

[–] recursive 4 hours ago

It's finally happening. It's almost a relief.

[–] root_axis 5 hours ago

Current flavor of the week variant of OpenClaw "feels useful". Now that's an endorsement.

[–] Genego 4 hours ago

I did try OpenClaw myself when it came out. But I felt like there was nothing really new there over what we had been building for the previous years; just consumer adoption happening. Definitely feels like this new wave of products (Muse, Duet, Dots, etc) is onto something. I have been them daily, feel like they are incredibly useful, and have clients ask to build something like this for them (they never asked for OpenClaw'esque products).

[–] dovin 6 hours ago

Just got access today. I like the channels, but I'm actually finding a few things frustrating compared to Dots: no phone calls (convenient to braindump and have it clean up later) and I don't have shell access to its underlying VM like with Dots. It's also weirdly hard / impossible to spawn Claude Code for Web workers and I can't connect the agent to a desktop and have it spawn jobs there. It doesn't seem all that faster than Dots.

It's early and obviously they'll keep iterating. It definitely feels less "make it yours" than Dots and that does feel intentional.

[–] TomGarden 7 hours ago

I disagree on the voice interaction point. I DO want voice modes to be glorified STT-TTS.

The OpenAI AVM is uncanny and over-eager to me. You can run Anthropic voice on Opus since a handful days and it's incredibly snappy without feeling weird.

[–] R_D_Olivaw 4 hours ago

Agreed. I don't want the extra bits that try to mimick humans.

Sometimes it's as though they added background noise to make it sound like it's looking for things. Or even the long "hmmm, let me cheeeckkkk...".

Also, I think I'm ready to be able to choose from a robotic voice.

[–] quixoticelixer- 5 hours ago

I would be much more tempted to switch over to this from Hermes if I didn't have to deal with claudese.

[–] brightball 3 hours ago

I will be shocked if anything can beat Grok Bot for usefulness at this point.

It took me 2 days to hit “how did I ever survive without this” and it’s only gotten better.

[–] hangrybear666 7 hours ago

Yeah sure. Let's talk again in 3 months once a new paradigm has been invented to further abstract and alienate us from the work we are doing.

[–] teekert 7 hours ago

Can't use it, won't log me in on GrapheneOS. ("Can't be verified via Google Play")

[–] aniceperson 7 hours ago

my rpi5 + wireguard + pi.dev is my 'duet'

[–] lazzlazzlazz 4 hours ago

All the features the author describes here are present in Grokbot, which I find the best by far, and I've tried Hermes Agent, OpenClaw, NanoClaw, IronClaw, Instinct, Muse, dots, and many others at length. The simplest and best overall package so far is Grokbot, by a respectable margin. Although I haven't tried Duet!

[–] tinyhouse 6 hours ago

Sounds like Duet is more like Grokbot than Muse? Haven't used it but the author is talking about channels, github policy access for specific channels, etc. That doesn't sound like the Muse use case.

I actually use Claude's agents on the Claude platform for all things work related and it works pretty well. It's not cheap though if you use the more expensive models and setting it up requires currently some initial effort for connecting integrations like Slack the average user is unlikely to do. It also doesn't let you change things. Every change requires a new deployment. As a cloud agent that runs on a schedule it's pretty good. I think they just need to figure out a single place for this. Currently you can set up agents on the claude platform, as part of claude code, and as part of regular claude.

[–] dgellow 7 hours ago

Sounds like things the competition will copy in a few days? Or am I missing something?

[–] dbbk 7 hours ago

What do you mean copy? This is the copy, it's already behind Muse, Dots

[–] bluecalm 7 hours ago

The killer feature is having the best model behind it. Not so easy to copy. Grok Bot tries to compete with a really good harness/infra and the team is doing a great job but having Grok as a model behind id is a huge disadvantage.

[–] fHr 6 hours ago

[–] Computer0 3 hours ago

I am not having good luck with it when using it the same way I used Open AI Dot. It will not even download from dropbox.

[–] imagetic 7 hours ago

I'm amazed anyone on HN is willing to run Muse that isn't a FB employee or tech journalist.

[–] meowface 7 hours ago

I don't trust Facebook or Zuckerberg, but I also don't really care about my own personal privacy from large corporations. I will use whatever is best/cheapest, generally. And I say this as someone who hates advertising and adtech.

[–] cautiouscat 7 hours ago

Or GrokBot that isn’t an Elon employee.

[–] enraged_camel 6 hours ago

I haven't used GrokBot, but I have been sort of amazed by how popular it is. I don't have the numbers but even my friends who otherwise dislike Elon have at least given it a try, and several of them use it daily.

It makes me want to check it out just to see if there's some magic UX juice they put in there. Then again I also don't want to take any risks heh.

[–] joshgachnang 6 hours ago

I've also noticed anything related to GrokBot gets massively boosted on X, which is probably serving as a huge marketing channel. Plus anyone with a Cursor subscription gets access.

[–] skinfaxi 4 hours ago

Cursor started to push it recently, too. They put a popup when I logged in to tell me about it.

[–] mcmcmc 6 hours ago

Most people probably wouldn’t cop to the real reason they use the CSAM and non-consensual nudes generator

[–] nl 6 hours ago

I know someone who isn't an Elon fan at all but has switched completely from Dot to Grokbot. Most of their company has done the same and now they are contemplating cancelling their OpenAI plans (they still use Claude heavily, but Duet isn't available on Teams plans).

[–] lavezzi 3 hours ago

Can you name the company to give the rest of us a heads up?

[–] HDThoreaun 6 hours ago

Ive never had a meta account. Zuckerberg still has all my data and knows everything about me. At some point people decide its not worth hiding from mega corps

[–] ipsum2 7 hours ago

That's because HN is an echo chamber. Plenty of tech people use and enjoy Muse/Instagram/Facebook, but the would be downvoted if they expressed it.

[–] baq 7 hours ago

I disagree. I would gladly hear why they think there are no privacy concerns or why they don’t care about their privacy so much. I find it hard to understand.

[–] nl 6 hours ago

I'm a long time HN user. I work on privacy tech (as in I'm paid to work on it).

I have a Facebook account and use Instagram.

My reasons are many:

I believe it should be my choice what happens to my data, and don't particularly care if Meta has the parts I share with them.

I don't believe advertising is intrinsically evil, and in fact I think it's probably been a net-good for society.

I think Meta is a better steward of my data than ISPs or telecoms.

[–] saikia81 4 hours ago

what about the abuse of advertisements to cause civil unrest, abusing data to control political elections, advertisements that break consumer laws. Facebook has shown it cares more for money than being a steward. Don't see how you can trust them.

[–] nl 16 minutes ago

I'm aware there are downsides.

I'm old enough to have lived without the internet. I'll take this over that anytime.

> Facebook has shown it cares more for money than being a steward. Don't see how you can trust them.

It's precisely this love of money that make me trust them. They value what they get from my data, so they do their best to keep it to themselves.

That's much better than ISP (for example) that sell raw traffic data.

[–] dovin 6 hours ago

I use Muse to read X for me with a burner account, since I want to follow certain topics but don't want to use X. It posts digests a couple times a week to a discord channel I gave it access to. It's fine and I think there are plenty of ways of using it that give essentially nothing to Meta in the way of data. Still I wouldn't recommend it to any of my non-technical friends, since I think the natural progression with these things is to just connect more and more parts of your life to them.

[–] Karrot_Kream 7 hours ago

Almost all of my tech worker friends use Instagram. It's almost universally used. I never talk about it because anything positive about any Meta property on this site gets downvoted. It's been like this for years and only gotten worse over time.

[–] BeetleB 6 hours ago

I said the same for years after smartphones came around.

In my book, if you use a Smartphone that's either Android or iPhone, I can't take seriously the claim that you value your privacy.

Privacy is a continuum.

[–] andix 8 hours ago

Note to everyone writing blog posts: Please start your post with one or two sentences explaining what you're posting about.

What is Duet? I have no idea even after skimming the article.

Edit: first google result is https://www.duetdisplay.com

[–] vineyardmike 8 hours ago

Well the start the article saying they're not affiliated with Anthropic, and then talk about "Duet", saying "anthropic has done it again". They then list a handful of products they've tried, including "a host of other personal-assistant-type harnesses".

Is it not obvious that this is an Anthropic-made "personal-assistant-type harnesses"?

https://claude.com/resources/articles/claude-duet

[–] andix 6 hours ago

I could deduct some of that from this sentence, sure.

But I wrote this comment because the article was frustrating for me to read. One single sentence would've made it much less confusing:

  Look, I really didn’t want to like **Claude** Duet, **Anthropic's new personal AI assistant**. I don’t want to be an Anthropic fanboy and ...

I'm also known to flag this kind of shortcoming in all technical documentation I review, and it annoys a lot of people that I do. I've put my judgement to the test already a few times in team meetings when someone objected this specific remark. Many times I could prove that there are always a few people in the room that would've benefited a lot from such a simple clarification. They just didn't speak up or ask, because they thought they were the only ones that didn't know a specific term.

My intention is also not to be rude, just creating some awareness for this issue.

[–] user- 8 hours ago

The first two sentences menthon anthropic so its this. https://claude.com/resources/articles/claude-duet

[–] andix 6 hours ago

I know. The full product's name seems to be "Claude Duet", which would make everything much clearer and easier to google it.

[–] longerthoughts 7 hours ago

Meh, I think it's fine for blog posts to assume familiarity like this if it reflects their niche. They're writing for the type of person who already knows what Duet is. It's not the NYT where there's a massive general audience.

In fact, the choice not to explain things tells me a lot about the shared knowledge of a niche when I'm looking in from the outside. I don't mind the extra digging to decode it and the implied familiarity gives me information about its importance.

[–] xnx 8 hours ago

[Anthropic Claude] Duet is the future of chatbot interaction

[–] hmokiguess 7 hours ago

Once the novelty wears off it goes in the drawer next to the rest of AI innovation

[–] hlynurd 7 hours ago

Are you talking about self-organising maps or the sci-fi level capability we've seen in the last year?

[–] thisisnotauser 8 hours ago

Every time I see an article about another agentic harness all I can think is "people still run other people's code?"

In all seriousness, the complaints here just feel like UI friction you can address yourself if you just build your own harness (e.g. snappy replies -> parallel threads and appending to common context).

[–] infecto 7 hours ago

It is just rsync after all right?

[–] gessha 7 hours ago

I’m on the same opinion. Why do I need to limit myself to something proprietary and in somebody else’s control when I can build and morph a tool to my needs.

The only reason I even look at frontier lab apps is for inspiration or shortcuts to my own implementation.

Using their apps feels like QA testing a trainwreck anyway with all of their UI/UX being an afterthought.

[–] arjie 5 hours ago

I do agree. We have a household assistant and it's not particularly different in functionality from the ones by mainstream labs. It's not a lot of work to build. Because we run on our own LLM base we can also do things like plumb our home cameras into it and so on.

[–] 9dev 7 hours ago

These kinds of replies always astound me. Like, I have a better use for my attention and token limit than building and fiddling with my custom agent harness?

[–] BeetleB 5 hours ago

Depending on what you need, you can build a custom harness via pi really fast. I tried OpenClaw for weeks, and it took me a lot less time building my harness than trying to figure OpenClaw out.

[–] scottyah 7 hours ago

I think it's more cron jobs, RESTful-type wait and react methods, and a harness that is calibrated to being a good assistant.

i.e. It saw I had an optometry-type appointment this morning and supplied some questions I might want to ask or consider.

[–] Rabbit504030201 4 hours ago

The first sentence really do reads like trolling.