mickdarling

6 ms·

1.
by mickdarling 25d ago
Literally the first thing I thought of. Is he trying to reference an Antichrist?
2.
by mickdarling 26d ago
I created MCP AQL, which is an extension to the MCP spec, specifically to reduce the bloat for MCP tools. It only has five CRUDE endpoint: Create, Read, Update, Delete, and Execute using a GraphQL-like structure for tool calling of the oper
3.
by mickdarling 29d ago
If it stores unique information, by definition it can store arbitrary information because it can point to arbitrary information. So they can have it relate to anything they want. Even a full breakdown of the original text if they choose.
4.
by mickdarling 1mo ago
Watermarks are context poisoning.
5.
by mickdarling 3mo ago
If only the AI tools didn't shut you down every time you were trying to red team your own tools. I've had to come up with all kinds of workaround scenarios, effectively bypassing the AI security processes in order to stress test
6.
by mickdarling 3mo ago
No, this is their get out of jail free card if people start complaining about the model being dumb or forgetful or lying, they can just say, oh well, you must have been doing something that triggered its distillation prevention technique. A
7.
by mickdarling 3mo ago
The tags are actually displayed in raw text not rendered.
8.
by mickdarling 3mo ago
Below is the EXACT text in Claude Desktop introducing Fable 5, including the very professional looking break tags, and at least I know where the links begin and end by looking at the anchor tag there. They obviously put their best model on
9.
by mickdarling 4mo ago
I effectively distill the frontier models by building whole sets of skills, personas, and other artifacts that I can then run on smaller models and get 10% even 20% improvements on models like haiku or local models. There's a lot of ro
10.
by mickdarling 5mo ago
I used it today to take a look at my previously built design system with Logos, branding, fonts, and everything else. After a lot of annoying tweaking back and forth, finally, I got something that was satisfactory. Then I looked at the usag
11.
Show HN: DollhouseMCP 2.0, open-source MCP composable AI building blocks(dollhousemcp.com)
3 points by mickdarling 5mo ago|0 comments
12.
by mickdarling 6mo ago
I'm working on a tool to determine which portions of an LLM process can be optimized, and how to measure that optimization and check whether it's optimizable at all. The shaping pattern that they talk about here is directly releva
13.
by mickdarling 6mo ago
Hey, LLM, take a look at these multiple hundred emails and docs in my docs folder from the last few years, before I started using AI, that I wrote personally. create a list of all of the idiosyncrasies that I have in my writing. Create a fi
14.
by mickdarling 6mo ago
Maybe for you reading a paper deeply is the most constructive way that you have to absorb information. For me, it is having a document and interrogating it. Maybe having many sets of documents about a whole category of information. Getting
15.
by mickdarling 6mo ago
Might be worth something to create an AI summary of the actual documents that are behind the curtain. That way the Client can have some assurance of what the deliverables are. It's a nice ability for the escrow service to be a trusted
16.
by mickdarling 6mo ago
I don't use Claw. It is way too dangerous. I built my own system where I know the ins and outs and how they can break. When it comes to agents' tasks, I tend to focus on things that I couldn't do before without automated agen
17.
by mickdarling 7mo ago
You could literally drop this into Claude Code or Codex and point it at a local fork of Zulip and have it build your bimodal version with triage and grazing styles.
18.
by mickdarling 7mo ago
I use an LLM behavior test to see if the semantic responses from LLMs using my MCP server match what I expect them to. This is beyond the regex tests, but to see if there's a semantic response that's appropriate. Sometimes the LLM
19.
by mickdarling 7mo ago
It looks like it does have an MCP Gateway https://github.com/github/gh-aw-mcpg so I may see how well it works with my MCP server. One of the components mine makes are agent elements with my own permissioning, security
20.
by mickdarling 7mo ago
This is where the desire to NOT anthropomorphize LLMs actually gets in the way. We have mechanisms for ensuring output from humans, and those are nothing like ensuring the output from a compiler. We have checks on people, we have whole indu
21.
by mickdarling 8mo ago
I'm looking at it right now as a tool I can hollow out and stuff in my own MCP server that also has personas, skills, an agentic loop, memory, all those pieces. I may even go simpler than that and simply take a look at it's gatewa
22.
by mickdarling 8mo ago
Thank you for the note. It's not a site I used all that often. Whether you had anything to do with it or not, I have no idea. And, since you didn't follow best practices and tell me directly rather than trying to score points her
23.
by mickdarling 8mo ago
Wasn't my project to manage. That was a consulting gig. And I fired the client right after this.
24.
by mickdarling 8mo ago
Had humans not been doing this already, I would have walked into Samsung with the demo application that was working an hour before my meeting, rather than the android app that could only show me the opening logo. There are a lot of really b
25.
by mickdarling 8mo ago
I think Claude Cowork should come with a requirement or a very heavily structured wizard process to ensure the machine has something like a Time Machine backup or other backups that are done regularly, before it is used by folks. The failur
26.
by mickdarling 8mo ago
This looks cool! And, to add to the list of shameless plugs for OSS markdown editors/renderers With mermaid support, I will add mine to the list: https://merview.com with full source code at https://github.com&#x
27.
by mickdarling 8mo ago
It certainly doesn't seem to have a trouble creating MIT licenses, that's for sure. I've had it insert an MIT license against my express direction instead of the AGPL license.
28.
Claude Code Unable to generate a AGPLv3 license due to content filtering policy(github.com)
7 points by mickdarling 8mo ago|3 comments
29.
by mickdarling 8mo ago
Claude code refuses to generate AGPL 3.0 licenses or modify them. I've run into this with several projects.
30.
by mickdarling 9mo ago
That's a fascinating path forward and most frustrating of all, if I understand this right, this was something that could have been discovered 20 years ago.

More ›