eigenvalue

8 ms·

61.
Show HN: The Ultimate MCP Client(github.com)
7 points by eigenvalue 1y ago|1 comments
62.
by eigenvalue 1y ago
I think that’s because it’s a Chinese Project. Same thing with Ant Design Components, which are really awesome but not as well known as they should be.
63.
by eigenvalue 2y ago
You’re right that it’s analogous in concept, but strategy distillation happens at a higher level: it encodes and transfers successful latent reasoning patterns as reusable “strategies,” without necessarily requiring direct gradient updates
64.
by eigenvalue 2y ago
I can see where you’re coming from, but not really. Unlike an RNN, the main transformer still processes sequences non-recurrently. The “sidecar” model just encodes internal activations into compressed latent states, allowing introspection a
65.
by eigenvalue 2y ago
I can assure you it’s not a joke. Compute power is increasing at a ridiculous pace, and highly capable models are getting smaller and smaller, now at the 30b parameter size and under. So even if it wouldn’t be pragmatic now, it could become
66.
by eigenvalue 2y ago
Interesting, it certainly wouldn’t take up much additional space, but I wonder if it would have any real impact, since it seems somewhat orthogonal to finding a faithful low-dimensional encoding of the activations.
67.
by eigenvalue 2y ago
Welcome Tom! Thanks for helping to make this the highest signal-to-noise ratio forum in the general technology/business space.
68.
Real-Time Introspective Compression for Transformers(github.com)
14 points by eigenvalue 2y ago|11 comments
69.
by eigenvalue 2y ago
I recently started thinking about what a shame it is that LLMs have no way of directly accessing their own internal states, and how potentially useful that would be if they could. One thing led to the next, and I ended up developing those i
70.
by eigenvalue 2y ago
This is great, I was debating whether I should do my latest project using the new OpenAI Responses API (optimized for agent workflows) or using MCP, but now it seems even more obvious that MCP is the way to go. I was able to make a pretty c
71.
by eigenvalue 2y ago
I used to use McFly too, but atuin is much better. Worth trying, takes one minute to install it.
72.
LLM Multi-Round Coding Tournament(github.com)
2 points by eigenvalue 2y ago|0 comments
73.
Show HN: LLM-Tournament – Have 4 Frontier Models Duke It Out over 5 Rounds(github.com)
1 points by eigenvalue 2y ago|0 comments
74.
Dr. GPT: Empowering Your Healthcare Decisions with AI(fixmydocuments.com)
1 points by eigenvalue 2y ago|1 comments
75.
by eigenvalue 2y ago
I wrote this article detailing my recent experience helping my dad navigate some complex health issues with AI models. In the process, I came up with a good set of prompts and procedures for getting really high-quality diagnostic feedback a
76.
TaxGPT: How to Use AI to Improve Your Tax Preparation and Lower Your Costs(fixmydocuments.com)
3 points by eigenvalue 2y ago|0 comments
77.
by eigenvalue 2y ago
I feel like humans don’t have great intuitions around the ideal format for an LLM. But an LLM probably does have a good sense for what would work well for itself. And also I think that we shouldn’t necessarily expect to find LLM targeted do
78.
Show HN: LLM-docs, software documentation intended for consumption by LLMs(github.com)
28 points by eigenvalue 2y ago|7 comments
79.
by eigenvalue 2y ago
Was it the word “the” instead of “a”? Trying to think what else it could even be.
80.
by eigenvalue 2y ago
Author here. I originally wrote the code for this project as part of my commercial webapp project (FixMyDocuments.com), but I was so intellectually stimulated by the creation of this code, and so excited by the new design possibilities it o
81.
Show HN: Recursive Descent Mindmap Generator(github.com)
2 points by eigenvalue 2y ago|1 comments
82.
by eigenvalue 2y ago
I don’t really play any games but the game play demo video looked like something that might entertain my 4 and 1 year olds for a few minutes at least (just watching, I doubt they could really play it), so I bought it. Congrats on your succe
83.
Show HN: Mindmap Generator – Marshalling LLMs for Hierarchical Document Analysis(fixmydocuments.com)
3 points by eigenvalue 2y ago|0 comments
84.
by eigenvalue 2y ago
Nice, probably saved a bunch of FANG devs a lot of hours of work trying to knock this off.
85.
Show HN: Analyzing PPP Loan Fraud with Advanced Python Data Analysis(github.com)
1 points by eigenvalue 2y ago|0 comments
86.
by eigenvalue 2y ago
I made a site recently that works pretty well for this for a lot of sample scanned PDFs I tried, you might get good results: https://fixmydocuments.com/ I also made a simple iOS app that basically just uses the built in OCR
87.
by eigenvalue 2y ago
If you want to be able to select certain files quickly and visually, and work with private repos or just local files, try this open source tool I made: https://github.com/Dicklesworthstone/your-source-to-prompt.h...
88.
by eigenvalue 2y ago
I haven’t, but I bet it would work pretty well, particularly if you tweaked the prompts to explain that it’s dealing with Ancient Greek or whatever and give a couple examples of how to handle things.
89.
by eigenvalue 2y ago
I think the current sweet-spot for speed/efficiency/accuracy is to use Tesseract in combination with an LLM to fix any errors and to improve formatting, as in my open source project which has been shared before as a Show HN: http
90.
by eigenvalue 2y ago
I basically wrapped this in a simple iOS app that can take a PDF, turn it into images, and applies the native OCR to the images. It works shockingly well: https://apps.apple.com/us/app/super-pdf-ocr/id64796742

More ›