supermdguy

8 ms·

61.
by supermdguy 7mo ago
Probably referencing this: https://news.ycombinator.com/item?id=47034087
62.
by supermdguy 7mo ago
Yeah, OpenClaw agents have a full set of tools to interact with a browser in arbitrary ways. My idea was to instead give it a tool for a browser wrapper with a limited API surface. And that tool could use LLMs internally in specific context
63.
by supermdguy 7mo ago
I like simonw's definition: "An LLM agent runs tools in a loop to achieve a goal." I guess agent isn't the best term here since the LLM wouldn't be driving the logic in the daemon. Using an LLM to select which item
64.
by supermdguy 7mo ago
One promising direction is building abstraction layers to sandbox individual tools, even those that don't have an API already. For example, you could build/vibe code a daemon that takes RPC calls to open Amazon in a browser, searc
65.
by supermdguy 8mo ago
I think there's still value in building quality products, but AI makes it easy to build something that appears good but doesn't actually work that well. It's very difficult to communicate the thought and intentionality that w
66.
by supermdguy 8mo ago
Has anyone had success using skills like these without an agent that supports them? I’m using the Zed agent, which doesn’t have skills support, and was thinking of just adding in a summary of the skills directory and how to use it inside my
67.
by supermdguy 9mo ago
Looks like this would affect around 4.3% of chats (the "Self-Expression" category from this report[0]). Considering ChatGPT's userbase, that's an extremely large number of people, but less significant than I thought base
68.
by supermdguy 9mo ago
That's a good point, thinking about it some more, I think the business logic feels so trivial that it would make the code harder to reason about if it were separated from the effects. Currently, I have one giant function that pulls dat
69.
by supermdguy 9mo ago
Typescript, using Zod with Express for parameter validation.
70.
by supermdguy 9mo ago
> A common pattern would be to separate pure business logic from data fetching/writing. So instead of intertwining database calls with computation, you split into three separate phases: fetch, compute, store (a tiny ETL). First fetc
71.
by supermdguy 9mo ago
How do you guys share types between your frontend and backend? I've looked into tRPC, but don't like having to use their RPC system.
72.
by supermdguy 9mo ago
Reading the code, I was surprised to see that cd was implemented by calling out to the os library. I assumed that was something the shell or at least userspace handled. At what level does the concept of a “current directory” exist?
73.
by supermdguy 10mo ago
If your output schema doesn’t capture all correct outputs, that’s a problem with your schema, not the LLM. A human using a data entry tool would run into the wrong issue. Letting the LLM output whatever it wants just makes it so you have to
74.
Micromort(en.wikipedia.org)
1 points by supermdguy 11mo ago|2 comments
75.
by supermdguy 11mo ago
Really like the philosophy, and the UI looks clean. You mentioned grammar briefly, but I’m curious if you think that’s also a component that could be learned through the app? One thing that’s nice about Duolingo (despite its flaws) is that
76.
First In-Human Trial of CRISPR Shown to Safely Lower Cholesterol(newsroom.clevelandclinic.org)
3 points by supermdguy 11mo ago|0 comments
77.
by supermdguy 1y ago
That corresponds to a 10/15, which is actually really good (median is around 6) https://artofproblemsolving.com/wiki/index.php/AMC_historica...
78.
by supermdguy 1y ago
Interesting work. Not super familiar with neural architecture search, but how do they ensure they’re not overfitting to the test set? Seems like they’re evaluating each model on the test set, and using that to direct future evolution. I get
79.
by supermdguy 1y ago
Just guessing, but the new Opus was probably RL tuned to work better with Claude Code's tool calls
80.
by supermdguy 1y ago
How does your tool library work? Who organizes it? Sounds really interesting.
81.
Testing the Hard Stuff and Staying Sane [video] (2014)(youtube.com)
1 points by supermdguy 1y ago|0 comments
82.
by supermdguy 1y ago
Most providers will just end the chat if it reaches the max context window.
83.
The most-cited papers of the twenty-first century(nature.com)
6 points by supermdguy 1y ago|2 comments
84.
Hastert Rule(en.wikipedia.org)
9 points by supermdguy 2y ago|1 comments
85.
The Indispensable Opposition (1939) [pdf](cdn.theatlantic.com)
2 points by supermdguy 2y ago|0 comments
86.
Agent Interoperability(humansimulation.ai)
3 points by supermdguy 2y ago|0 comments
87.
by supermdguy 2y ago
clippings.io has a browser extension that scrapes all your highlights from Amazon's website and lets you download them in various formats.
88.
Reading the Mind in the Eyes(s3.amazonaws.com)
1 points by supermdguy 2y ago|0 comments
89.
Discord obliterated a YouTube view count record. It may have been an accident(mashable.com)
10 points by supermdguy 3y ago|2 comments
90.
by supermdguy 3y ago
Code is available! They have a few different discovered solutions. https://github.com/google-deepmind/funsearch

More ›