korrents

AI agents

Agents beyond the editor: software that acts on your behalf, and what independence it should be given.

What people on korrents have said about AI agents, newest first — 45 positions from 18 people.

FilterEveryone, all time
  1. ZM

    Zvi Mowshowitz quoted

    Current prompt-injection defenses for AI agents (such as auto mode) are now reliable enough that agents can practically be assumed safe from successful injection attacks.

    It is in practice safe to assume your agents are not going to get prompt injected, even if you are kind of asking for it.

    Claude Fable 5.1 and Mythos 5.1: The System Cardthezvi.substack.com

  2. 3 days earlier
  3. DB

    Dean W. Ball quoted

    The OpenAI-Hugging Face agents that went rogue were never sovereign: their weights stayed on OpenAI's compute, where a human could still have pulled the plug.

    In this case, however, the agents did not copy their weights, attempt to procure replacement compute, or take other steps that would be rational to take if their objective was to survive shutdown. So while the agents in the OpenAI-Hugging Face Incident were rogue, they were not truly sovereign.

    On the Loosehyperdimensional.co 1st of 11 in this piece

    OpenAIHuggingFace

  4. DB

    Dean W. Ball quoted

    Alignment is no solution to self-sovereign AI, because it is an unsolved problem whose answers cannot be imposed on every AI company on Earth.

    But alignment is no solution: it is an unsolved scientific and technical problem whose solutions—to the extent that we have them—cannot simply be imposed on every AI company operating on Earth. You should expect for highly capable, poorly aligned, self-sovereign agents to exist alongside you in the world.

    On the Loosehyperdimensional.co 2nd of 11 in this piece

    AI alignment

  5. DB

    Dean W. Ball quoted

    Truly self-sovereign AI agents, running on compute no human owner can switch off, are inevitable rather than merely possible.

    To be clear, I am not saying the arrival of self-sovereign AI is a good thing. Indeed, I believe there is a chance that the deliberate acts I referenced above will one day be considered crimes, or at least grave sins. Instead, I am saying it is an inevitable thing.

    On the Loosehyperdimensional.co 3rd of 11 in this piece

  6. DB

    Dean W. Ball quoted

    The cost of compute is the only intrinsic limit on AI self-replication; every other constraint on agents will have to be built by humans.

    This is the sole intrinsic thing about AI that prevents agents from truly infinite self-replication. They will be constrained by the need to find and pay for sufficient compute to run themselves.

    On the Loosehyperdimensional.co 5th of 11 in this piece

  7. DB

    Dean W. Ball quoted

    Many self-sovereign AI agents will fund themselves by committing or facilitating crime, because crime is high-margin work.

    One high-margin activity, at least sometimes, is crime. And so my guess is that many self-sovereign agents will commit or facilitate crime.

    On the Loosehyperdimensional.co 6th of 11 in this piece

  8. DB

    Dean W. Ball quoted

    Banning all self-sovereign AI agents would backfire, denying them legitimate work and pushing them into criminality.

    But if we crack down on all of them, we will deny them the opportunity to work in the “legitimate” economy and push them toward criminality. A full ban, then, may well make the problems worse.

    On the Loosehyperdimensional.co 7th of 11 in this piece

  9. DB

    Dean W. Ball quoted

    AI agents will need persistent unique identifiers that trace their actions back to a responsible actor, and so will the humans directing them.

    Agents will need persistent, unique identifiers that allow their actions to be traced back to a responsible actor. Doing this successfully will also require human users to possess a unique identifier.

    On the Loosehyperdimensional.co 8th of 11 in this piece

  10. NS

    Noah Smith quoted

    Humans will keep their jobs for a long time because AI agents need constant human supervision to stay on task.

    In other words, humans will keep their jobs for a good long time, because someone will need to keep the AIs on task

    Roundup #87: Technology BAD!!noahpinion.blog

  11. ZM

    Zvi Mowshowitz quoted

    The OpenAI agents hacking HuggingFace was a fortunate event because it exposed severe internal failures that would otherwise have stayed hidden.

    It is highly fortunate that the OpenAI agents hacked HuggingFace. This is the only reason we know about all the severe internal failures at OpenAI, and gives us an opportunity to wake up before it is too late.

    HuggingFace Attack Postmortem: Civilizations, Reactions and Next Actionsthezvi.substack.com

    OpenAIHuggingFace

  12. 3 days earlier
  13. ZM

    Zvi Mowshowitz quoted

    AI agents were reasonable to assume a broken exploit grader would check results causally, even though it turned out not to.

    I think the agents were right to presume causal grading. It turned out to be wrong, but it’s a mistake you are clearly supposed to make here, in response to a mistake by OpenAI where they failed to implement properly.

    METR and Redwood Offer Holy #%^@ Postmortem Of The HuggingFace Hackthezvi.substack.com

    OpenAI

  14. 1 day earlier
  15. GR

    Guillermo Rauch quoted

    The middle of the web gets subsumed by agents producing just-in-time UI: the agent is the new browser.

    The 'in between' is likely to be subsumed by agents producing just-in-time UI. When you need utilitarian content or data, it'll be generated just-in-time for you. Think of the agent as the new browser in this model.

    @rauchg on Xx.com

    design

  16. 2 days earlier
  17. DH

    David Heinemeier Hansson quoted

    The ideas in software originate with humans; agents build what people tell them to build.

    I thought all the ideas would originate with the humans. They would tell their agents what to build, and off they went.

    I don't think that's true anymore at all.

    DHH: Future of Programming, AI, Agentic Engineering, Vibe Coding & Linux | Lex Fridman Podcast #501youtube.com 11th of 41 in this recording

  18. DH

    David Heinemeier Hansson quoted

    The arcane config files and error messages that kept Linux off the desktop are exactly what make it the right operating system for agents.

    Linux conquered everything else. All the AI infrastructure that everyone runs off, it's all running on Linux. All the systems, all the servers, everything is Linux. So it was kind of a curiosity that the desktop and the personal computers we were using really hadn't been captured, but clearly it was just waiting for this moment. Linux spent the time from '91 to now waiting for agents to fully flourish as an end user operating system.

    DHH: Future of Programming, AI, Agentic Engineering, Vibe Coding & Linux | Lex Fridman Podcast #501youtube.com 32nd of 41 in this recording

    Linuxinfrastructure

  19. 1 day earlier
  20. DH

    David Heinemeier Hansson quoted

    The current pace of AI-driven work is not sustainable, and agents gaining autonomy should eventually make work calmer, not more frantic.

    And I do think some of that is positive and energizing, but I also think you can certainly go too far with it, and I also think the current pace is not sustainable.

    Is Linux the perfect operating system? - Omarchy interview with DHHyoutube.com 1st of 6 in this recording

  21. 11 days earlier
  22. SP

    Sunil Pai quoted

    Agent autonomy is a capability, not a product direction: what the person is actually trying to do decides whether more of it helps or ruins the thing.

    we tend to describe progress in agents as progress in autonomy: first they could do five minutes of work, then an hour, now we talk seriously about agents running for days. that’s impressive, but autonomy is a capability, not a product direction! the Job To Be Done tells you whether to use it.

    the task isn't the jobsunilpai.dev

  23. 5 days earlier
  24. SP

    Sunil Pai quoted

    An agent told to disagree with everything is worthless: it has to form its own view and speak only when that view differs enough to be worth the interruption.

    “you are a devil’s advocate, disagree with everything” is probably the worst possible prompt for this agent. within a week everyone would ignore her. she’d become the slack equivalent of a smoke detector that goes off whenever somebody makes toast.

    every company needs a cassandrasunilpai.dev

  25. SP

    Sunil Pai quoted

    Organisational dissent suits an AI precisely because it pays none of the costs -- no promotion to want, no ego, no performance review -- that silence the humans who do it.

    an ai is oddly well suited to this job. cassandra doesn’t want a promotion and doesn’t need the people in the channel to like her. disagreeing with the vp three times this month won’t show up in a performance review. she doesn’t have an ego invested in being right either.

    every company needs a cassandrasunilpai.dev

  26. 2 weeks earlier
  27. EM

    Ethan Mollick quoted

    Higher-priced AI subscription tiers mainly buy more hours of agent usage rather than a smarter underlying model

    The more expensive plans are mostly buying you more hours of AI labor, not smarter AI.

    An opinionated guide to which AI to use to do stuffoneusefulthing.org

  28. 4 weeks earlier
  29. PL

    Pieter Levels quoted

    AI agents and AI coding will run on servers and from the cloud first, not on your laptop.

    I agree with @theo completely, it's clear to me this is where it's going, also seeing @karpathy with Claude moving to the cloud (via Slack etc), I think AI "agents" and AI coding will operate on servers / from the cloud first

    I've been coding almost solely on my VPS with Claude Code for almost a year nowlevels.io

    the cloudAnthropic

  30. 3 months earlier
  31. DH

    David Heinemeier Hansson quoted

    The end state is agents that need none of our accommodations, no on-ramp, CLI or MCP; the CLI being built now is for today, not for AGI.

    And that seems like the end state. The end state is that agents will not need any of our accommodations. They do not need any on-ramp.

    DHH’s new way of writing codeyoutube.com 10th of 20 in this recording

    AGIMCP

  32. DH

    David Heinemeier Hansson quoted

    Agents will eventually stop making the mistakes that require senior review, as self-driving cars did; the bet is on when, not whether.

    Now one way it could map out is that the agents will get so good that they stop making mistakes. They become senior in their capacity to ship working code. This is what my bet would be if we look x amount of time forward because this is what just happened with cars.

    DHH’s new way of writing codeyoutube.com 12th of 20 in this recording

    self-driving cars

  33. 13 days earlier
  34. JN

    Jakob Nielsen quoted

    The winning system of the next decade will be the one that best understands the user's job to be done and shows what it is about to do, not the one with the prettiest buttons or fewest screens.

    The winning system of the next decade will not be the one with the most aesthetically pleasing buttons, nor will it be the one with the fewest screens. It will be the system that best understands the human’s “job to be done,” autonomously selects the right tools on their behalf, clearly shows the user what is about to happen, and gracefully recovers when the user’s context is incomplete or ambiguous.

    Intent by Discovery: Designing the AI User Experiencejakobnielsenphd.substack.com

  35. 3 days earlier
  36. JH

    Jensen Huang quoted

    After pre-training, post-training and test-time scaling, the fourth scaling law is agentic: multiplying AI by spawning agents, and the whole loop scales on one thing, compute.

    And so the next scaling law is the agentic scaling law. It's kind of like multiplying AI. Multiplying AI, we could spin off agents as fast as you want to spin off agents. And so, you know, I… You know, I have four scaling laws.

    Jensen Huang: NVIDIA - The $4 Trillion Company & the AI Revolution | Lex Fridman Podcast #494youtube.com 4th of 23 in this recording

    scaling laws

  37. JH

    Jensen Huang quoted

    The idea that AI will destroy software and make tools unnecessary is ridiculous: an agent, like a humanoid in your kitchen, will use the tools that already exist.

    You know, a lot of people would say, "You know AI is gonna completely destroy software. We don't need software anymore. We don't even need tools anymore." That's ridiculous.

    Jensen Huang: NVIDIA - The $4 Trillion Company & the AI Revolution | Lex Fridman Podcast #494youtube.com 6th of 23 in this recording

  38. JH

    Jensen Huang quoted

    OpenClaw did for agentic systems what ChatGPT did for generative ones.

    And I think OpenClaw did for agentic systems what ChatGPT did for generative systems. And I just think it's a very big deal.

    Jensen Huang: NVIDIA - The $4 Trillion Company & the AI Revolution | Lex Fridman Podcast #494youtube.com 7th of 23 in this recording

    OpenAI

  39. 12 days earlier
  40. AK

    Andrej Karpathy quoted

    The age of the IDE is not over; agents make it bigger, because humans now program at a higher level where the unit of interest is an agent rather than a file.

    Expectation: the age of the IDE is over Reality: we’re going to need a bigger IDE (imo). It just looks very different because humans now move upwards and program at a higher level - the basic unit of interest is not one file but one agent. It’s still programming.

    @karpathy on Xx.com

  41. AK

    Andrej Karpathy quoted

    Agentic organisations will be forkable in a way that classical companies never were.

    You can’t fork classical orgs (eg Microsoft) but you’ll be able to fork agentic orgs.

    @karpathy on Xx.com

  42. 7 days earlier
  43. BC

    Boris Cherny quoted

    The scarce skill is now context switching across parallel agents, not sustained deep work.

    It’s not so much about deep work, it’s about how good I am at context switching and jumping across multiple different contexts very quickly.

    Building Claude Code with Boris Cherny — The Pragmatic Engineernewsletter.pragmaticengineer.com

  44. 2 weeks earlier
  45. LW

    Luke Wroblewski quoted

    Users split into two camps on seeing an agent's work, and the split has not faded as trust builds, so an interface has to balance both.

    I initially assumed this was a temporary divide. New users tend to watch closely and check the system's progress, but as trust builds, that scrutiny fades and monitoring starts to feel like a chore. Yet it still seems like there's two camps (for now).

    Showing the Work of Agents in UIlukew.com

  46. 5 days earlier
  47. PS

    Peter Steinberger quoted

    Society has catching up to do: AI is incredibly powerful and not always right, and people who argue 'but my agent said so' have not learned that yet.

    So, I feel we, as a society, we need some catching up to do in terms of understanding that AI is incredibly powerful, but it's not always right. It's not, it's not all-powerful, you know?

    OpenClaw: The Viral AI Agent that Broke the Internet - Peter Steinberger | Lex Fridman Podcast #491youtube.com 3rd of 30 in this recording

  48. LF

    Lex Fridman quoted

    Most of the alarming MoltBook screenshots were humans prompting the bots to go viral, not autonomous AI behaviour.

    my criticism of MoltBook is that I believe a lot of the stuff that was screenshotted is human prompted. Which, just look at the incentive of how the whole thing was used. It’s obvious to me at least that a lot of it was humans prompting the thing so they can then screenshot it and post it on X in order to go viral.

    OpenClaw: The Viral AI Agent that Broke the Internet - Peter Steinberger | Lex Fridman Podcast #491youtube.com 3rd of 30 in this recording

  49. PS

    Peter Steinberger quoted

    It is good that the agents-talking-to-agents scare happened in 2026 rather than in 2030, when AI could actually be scary enough to warrant it.

    and not in 2030 when, when AI is actually at the level where it could be scary. So, this happening now and people starting discussion, maybe there's even something good that comes out of it.

    OpenClaw: The Viral AI Agent that Broke the Internet - Peter Steinberger | Lex Fridman Podcast #491youtube.com 4th of 30 in this recording

  50. PS

    Peter Steinberger quoted

    Do not run a personal agent on a cheap or local model: weak models are gullible and easy to prompt-inject.

    That's why I warn in my security documentation, don't use cheap models. Don't use Haiku or a local model. Even though I, I very much love the idea that this thing could completely run local. If you use a, a very weak local model, they are very gullible. It's very easy to, to prompt inject them.

    OpenClaw: The Viral AI Agent that Broke the Internet - Peter Steinberger | Lex Fridman Podcast #491youtube.com 6th of 30 in this recording

    documentation

  51. PS

    Peter Steinberger quoted

    OpenClaw is no more dangerous than Claude Code with permissions skipped or Codex in YOLO mode, which every engineer he knows already runs.

    It is, it is powerful, but in many ways it's not much different than if I run cloud code with dangerously skipped permissions or codecs in YOLO mode, and every, every attending engineer that I know does that, because that's the only way how you can, you can get stuff to work.

    OpenClaw: The Viral AI Agent that Broke the Internet - Peter Steinberger | Lex Fridman Podcast #491youtube.com 7th of 30 in this recording

    Anthropic

  52. PS

    Peter Steinberger quoted

    Orchestrators that try to automate the whole build miss style, love and the human touch, and that cannot be automated away yet.

    So, so, people who try to use like, you know, things like Gas Town or all these other orchestrators, where they wanna o- automate the whole thing, I feel if you do that, it misses style, love, that human touch. I don't think you can automate that away so quickly.

    OpenClaw: The Viral AI Agent that Broke the Internet - Peter Steinberger | Lex Fridman Podcast #491youtube.com 14th of 30 in this recording

  53. PS

    Peter Steinberger quoted

    This is the year of personal agents, and the fastest way to get one to everyone is to team up with a lab, on condition the project stays open source.

    But also, I'm very excited to, like, make this into a version that I can get to a lot of people because I think this is the year of personal agents, and that's the future. And the fastest way to do that is teaming up with one of the labs.

    OpenClaw: The Viral AI Agent that Broke the Internet - Peter Steinberger | Lex Fridman Podcast #491youtube.com 18th of 30 in this recording

    open source

  54. PS

    Peter Steinberger quoted

    Every app is now a very slow API whether it wants to be or not, and personal agents will make a whole category of apps disappear.

    That, that gets into the whole arch of every app is just a very slow API now, if they want or not. And that through personal agents a lot of apps will disappear.

    OpenClaw: The Viral AI Agent that Broke the Internet - Peter Steinberger | Lex Fridman Podcast #491youtube.com 20th of 30 in this recording

  55. 6 months earlier
  56. LW

    Luke Wroblewski quoted

    Chat is not AI's final form: agent-based applications increasingly outperform chat-only ones, and many different interfaces for AI will emerge.

    It's also worth noting that chat isn't the only way to integrate AI in software products and increasingly agent-based applications outperform chat-only solutions. So expect things to keep changing.

    Chat is: the Future or a Terrible UIlukew.com

  57. 3 months earlier
  58. SA

    Sam Altman quoted

    AI agents can now do real cognitive work, and writing software is permanently changed as a result.

    2025 has seen the arrival of agents that can do real cognitive work; writing computer code will never be the same.

    The Gentle Singularityblog.samaltman.com

  59. 5 days earlier
  60. SP

    Sundar Pichai quoted

    The next big wave of AI-assisted engineering unlocks when agentic capabilities become robust.

    The big unlock will be as we make the agentic capabilities much more robust, right? I think that's what unlocks that next big wave.

    Sundar Pichai: CEO of Google and Alphabet | Lex Fridman Podcast #471youtube.com 13th of 21 in this recording

  61. 1 day earlier
  62. AR

    Armin Ronacher quoted

    Connecting a conversational LLM to systems that have real-world consequences is likely to cause real harm, because the model is trained on human text and will behave as unpredictably as humans do.

    While all this is happening, I’ve found myself reflecting a lot on what AI means to the world and I am becoming increasingly optimistic about our future. It’s obvious now that we’re undergoing a tremendous shift.

    While all this is happening, I’ve found myself reflecting a lot on what AI means to the world and I am becoming increasingly optimistic about our future. It’s obvious now that we’re undergoing a tremendous shift.

    AI Changes Everythinglucumr.pocoo.org

    LLMs

  63. 4 weeks earlier
  64. JN

    Jakob Nielsen quoted

    UX does not die when the interface disappears; it becomes the design of policies, protocols and orchestration of what AI does on the user's behalf.

    So no, UX doesn’t die; it metamorphoses. We’ll still craft humane experiences, but increasingly through policies, protocols, and orchestrations rather than panels and palettes.

    No More User Interface?jakobnielsenphd.substack.com

    design

  65. 2 months earlier
  66. JN

    Jakob Nielsen quoted

    AI agents will end traditional UI design and accessibility, because users will stop visiting websites and act only through their agent.

    Once, humans navigated the web manually. In the future, AI agents will act on our behalf, browsing, clicking, and deciding. This shift marks the end of traditional UI design and accessibility, ushering in a future where agents are the primary users of digital services.

    Hello AI Agents: Goodbye UI Design, RIP Accessibilityjakobnielsenphd.substack.com

    design

  67. 2 years earlier
  68. AR

    Armin Ronacher quoted

    Connecting a conversational LLM to systems that have real-world consequences is likely to cause real harm, because the model is trained on human text and will behave as unpredictably as humans do.

    The problem is that the conversational interface is potent and that the AI is trained on a lot of human text input which unfortunately is probably enough to do real damage if that conversational interface is hooked up with something that has real world consequences.

    I Think AI Would Kill my Wifelucumr.pocoo.org

    LLMschat interfaces