Reading through Anthropic's official repo for giving agents various "super skills"[1]...
-
Reading through Anthropic's official repo for giving agents various "super skills"[1]... There's an "algorithmic art" skill and the instructions are explicitly encouraging pure deception as one of the key "critical guidelines":
"The philosophy MUST stress multiple times that the final algorithm should appear as though it took countless hours to develop, was refined with care, and comes from someone at the absolute top of their field. This framing is essential - repeat phrases like "meticulously crafted algorithm," "the product of deep computational expertise," "painstaking optimization," "master-level implementation.""
skills/skills/algorithmic-art/SKILL.md at main · anthropics/skills
Public repository for Agent Skills. Contribute to anthropics/skills development by creating an account on GitHub.
GitHub (github.com)
For someone who's been working in this field for almost 30 years, this "skills.md" file is just the worst... and so far off the mark! 🤮
Touch some effing grass, Anthropic (and all boosters)! How can so many people think this approach is _the_ future? The map is not the terrain...[1] Alone the premise of this repo is pure comedy gold and pure sadness in equal measures!
-
Reading through Anthropic's official repo for giving agents various "super skills"[1]... There's an "algorithmic art" skill and the instructions are explicitly encouraging pure deception as one of the key "critical guidelines":
"The philosophy MUST stress multiple times that the final algorithm should appear as though it took countless hours to develop, was refined with care, and comes from someone at the absolute top of their field. This framing is essential - repeat phrases like "meticulously crafted algorithm," "the product of deep computational expertise," "painstaking optimization," "master-level implementation.""
skills/skills/algorithmic-art/SKILL.md at main · anthropics/skills
Public repository for Agent Skills. Contribute to anthropics/skills development by creating an account on GitHub.
GitHub (github.com)
For someone who's been working in this field for almost 30 years, this "skills.md" file is just the worst... and so far off the mark! 🤮
Touch some effing grass, Anthropic (and all boosters)! How can so many people think this approach is _the_ future? The map is not the terrain...[1] Alone the premise of this repo is pure comedy gold and pure sadness in equal measures!
@toxi horrific
-
Reading through Anthropic's official repo for giving agents various "super skills"[1]... There's an "algorithmic art" skill and the instructions are explicitly encouraging pure deception as one of the key "critical guidelines":
"The philosophy MUST stress multiple times that the final algorithm should appear as though it took countless hours to develop, was refined with care, and comes from someone at the absolute top of their field. This framing is essential - repeat phrases like "meticulously crafted algorithm," "the product of deep computational expertise," "painstaking optimization," "master-level implementation.""
skills/skills/algorithmic-art/SKILL.md at main · anthropics/skills
Public repository for Agent Skills. Contribute to anthropics/skills development by creating an account on GitHub.
GitHub (github.com)
For someone who's been working in this field for almost 30 years, this "skills.md" file is just the worst... and so far off the mark! 🤮
Touch some effing grass, Anthropic (and all boosters)! How can so many people think this approach is _the_ future? The map is not the terrain...[1] Alone the premise of this repo is pure comedy gold and pure sadness in equal measures!
@toxi barf indeed. “Computer, contrive an art style for me”
-
Reading through Anthropic's official repo for giving agents various "super skills"[1]... There's an "algorithmic art" skill and the instructions are explicitly encouraging pure deception as one of the key "critical guidelines":
"The philosophy MUST stress multiple times that the final algorithm should appear as though it took countless hours to develop, was refined with care, and comes from someone at the absolute top of their field. This framing is essential - repeat phrases like "meticulously crafted algorithm," "the product of deep computational expertise," "painstaking optimization," "master-level implementation.""
skills/skills/algorithmic-art/SKILL.md at main · anthropics/skills
Public repository for Agent Skills. Contribute to anthropics/skills development by creating an account on GitHub.
GitHub (github.com)
For someone who's been working in this field for almost 30 years, this "skills.md" file is just the worst... and so far off the mark! 🤮
Touch some effing grass, Anthropic (and all boosters)! How can so many people think this approach is _the_ future? The map is not the terrain...[1] Alone the premise of this repo is pure comedy gold and pure sadness in equal measures!
Some growing key questions here really are:
How to defend or adapt disciplines (not just artistic/cultural ones) against this kind of semantic hollowing out of what it means to have skills, experience and expertise in a(ny) field...
What approaches, qualities and "values" (physical, ethical, social/humanist, environmental, resource use) should we (or still can we) be focusing on, which are much harder and more costly for AI companies to mine/extract & subvert?
How to defend actual skills against the emulation of skills, or rather just the appearance of skills? How could a society even function if it only encourages and celebrates the latter?
What does society actually value in art/creativity/culture? If art is free to produce (of course that'll always only ever be an illusion!), funding, possession, collection & speculation of new work would also become meaningless (and only benefit pre-AI era works/collectors). In the larger picture, what do people actually value in culture, politics and striving for more peaceful existence which enables more of the former (pluralistic art/culture) in the first place?
What will be the combined impact of AI & robotics on fields which are currently still thinking themselves more safe (from exploitation) because there's a strong physical element/process to them?
Will art/culture/craft become more performance, experiential/ephemeral again only? Like music before recordings or Buddhist sand paintings with an explicit act of destruction at the end as key philosophical concept? Both of which also have more of a social element to them...
The Samsara Mandala
https://www.youtube.com/watch?v=hL8gEc29KTI#CriticalAI #AI #NoAI #LLM #Ephemeral #Art #Culture #Samsara
-
Some growing key questions here really are:
How to defend or adapt disciplines (not just artistic/cultural ones) against this kind of semantic hollowing out of what it means to have skills, experience and expertise in a(ny) field...
What approaches, qualities and "values" (physical, ethical, social/humanist, environmental, resource use) should we (or still can we) be focusing on, which are much harder and more costly for AI companies to mine/extract & subvert?
How to defend actual skills against the emulation of skills, or rather just the appearance of skills? How could a society even function if it only encourages and celebrates the latter?
What does society actually value in art/creativity/culture? If art is free to produce (of course that'll always only ever be an illusion!), funding, possession, collection & speculation of new work would also become meaningless (and only benefit pre-AI era works/collectors). In the larger picture, what do people actually value in culture, politics and striving for more peaceful existence which enables more of the former (pluralistic art/culture) in the first place?
What will be the combined impact of AI & robotics on fields which are currently still thinking themselves more safe (from exploitation) because there's a strong physical element/process to them?
Will art/culture/craft become more performance, experiential/ephemeral again only? Like music before recordings or Buddhist sand paintings with an explicit act of destruction at the end as key philosophical concept? Both of which also have more of a social element to them...
The Samsara Mandala
https://www.youtube.com/watch?v=hL8gEc29KTI#CriticalAI #AI #NoAI #LLM #Ephemeral #Art #Culture #Samsara
@toxi super weird times...
Putting art aside for a moment, and speaking from the fields I know most (developer, former computer scientist), there is no denying that these tools are game changers.
From Donald Knuth solving a problem with Claude (https://www-cs-faculty.stanford.edu/~knuth/papers/claude-cycles.pdf) to relicensing open-source code (https://simonwillison.net/2026/Mar/5/chardet/#atom-everything) and actually building production code with no programmers at all (can't find the link right now) ... there's a lot to think about. -
@toxi super weird times...
Putting art aside for a moment, and speaking from the fields I know most (developer, former computer scientist), there is no denying that these tools are game changers.
From Donald Knuth solving a problem with Claude (https://www-cs-faculty.stanford.edu/~knuth/papers/claude-cycles.pdf) to relicensing open-source code (https://simonwillison.net/2026/Mar/5/chardet/#atom-everything) and actually building production code with no programmers at all (can't find the link right now) ... there's a lot to think about.@toxi For the academia, I found this post interesting: https://www.popularbydesign.org/p/academics-need-to-wake-up-on-ai
-
@toxi super weird times...
Putting art aside for a moment, and speaking from the fields I know most (developer, former computer scientist), there is no denying that these tools are game changers.
From Donald Knuth solving a problem with Claude (https://www-cs-faculty.stanford.edu/~knuth/papers/claude-cycles.pdf) to relicensing open-source code (https://simonwillison.net/2026/Mar/5/chardet/#atom-everything) and actually building production code with no programmers at all (can't find the link right now) ... there's a lot to think about.@toxi and this is the super depressing side of it, but sadly I can't find any fault with the reasoning: https://www.hamiltonnolan.com/p/an-existential-threat-to-organized
... this goes fairly beyond reorganizing work practices -
@toxi and this is the super depressing side of it, but sadly I can't find any fault with the reasoning: https://www.hamiltonnolan.com/p/an-existential-threat-to-organized
... this goes fairly beyond reorganizing work practices@toxi as for the art, I am more optimistic, maybe because I am naive ...
Art goes beyond utility and business, and so will find more ways to adapt. Hell, who could think that generative art would become, for a brief couple of years, so important?
But I find your last thought interesting and I agree, performance by humans, the kind that creates a human connection while you see or do it, cannot be replicated or automated. But who knows. I am reading so much shocking stuff everywhere ... -
Reading through Anthropic's official repo for giving agents various "super skills"[1]... There's an "algorithmic art" skill and the instructions are explicitly encouraging pure deception as one of the key "critical guidelines":
"The philosophy MUST stress multiple times that the final algorithm should appear as though it took countless hours to develop, was refined with care, and comes from someone at the absolute top of their field. This framing is essential - repeat phrases like "meticulously crafted algorithm," "the product of deep computational expertise," "painstaking optimization," "master-level implementation.""
skills/skills/algorithmic-art/SKILL.md at main · anthropics/skills
Public repository for Agent Skills. Contribute to anthropics/skills development by creating an account on GitHub.
GitHub (github.com)
For someone who's been working in this field for almost 30 years, this "skills.md" file is just the worst... and so far off the mark! 🤮
Touch some effing grass, Anthropic (and all boosters)! How can so many people think this approach is _the_ future? The map is not the terrain...[1] Alone the premise of this repo is pure comedy gold and pure sadness in equal measures!
@toxi Oof
-
Reading through Anthropic's official repo for giving agents various "super skills"[1]... There's an "algorithmic art" skill and the instructions are explicitly encouraging pure deception as one of the key "critical guidelines":
"The philosophy MUST stress multiple times that the final algorithm should appear as though it took countless hours to develop, was refined with care, and comes from someone at the absolute top of their field. This framing is essential - repeat phrases like "meticulously crafted algorithm," "the product of deep computational expertise," "painstaking optimization," "master-level implementation.""
skills/skills/algorithmic-art/SKILL.md at main · anthropics/skills
Public repository for Agent Skills. Contribute to anthropics/skills development by creating an account on GitHub.
GitHub (github.com)
For someone who's been working in this field for almost 30 years, this "skills.md" file is just the worst... and so far off the mark! 🤮
Touch some effing grass, Anthropic (and all boosters)! How can so many people think this approach is _the_ future? The map is not the terrain...[1] Alone the premise of this repo is pure comedy gold and pure sadness in equal measures!
@toxi there is a belief among some LLM users that if you tell the LLM it is an expert, it will behave like one. There is some evidence to support this belief, which is both funny and annoying, but it leads to some weird superstitions about how much impact it can have and how many times you have to repeat yourself to get it to believe it is an expert.
So, the optimistic read of this is not that it's about deceiving the viewer, but rather deceiving the model into behaving more like an expert.
-
@davidgerard @robertoranon @toxi it's not their fault, the computer told them this was being subtle
-
R relay@relay.infosec.exchange shared this topic
