dmezzetti

7 ms·

1.
by dmezzetti 1mo ago
As an open source maintainer, I can chime in on this. With TxtAI, I've seen a large uptick in PRs ( https://github.com/neuml/txtai/pulls?q=is%3Apr+is%3Aclosed+s... ). While the extreme verbosity of Claude messa
2.
by dmezzetti 2mo ago
Very interesting project! Are there benchmarks for the "compatibility mode" or are all the numbers for the Gigatoken API?
3.
by dmezzetti 2mo ago
The recent abrupt removal of the ability to see who has starred a project isn't a good move. Things like this certainly erode trust. https://github.com/orgs/community/discussions/201209
4.
by dmezzetti 2mo ago
I agree that Kokoro is a good TTS model. If you're interested in an ONNX version and a permissively licensed TTS Tokenizer, I built a pipeline for that a while back: https://huggingface.co/NeuML/kokoro-base-onnx
5.
by dmezzetti 2mo ago
100% agree on this. I've been working on small local models for years with txtai ( https://github.com/neuml/txtai ). I've published close to 100 models that can run local for RAG, Agents, Vector Search and more
6.
by dmezzetti 2mo ago
Interesting project. Happy to see someone who shares an interest in tiny vector embeddings models. I've worked on tiny (1MB - 4MB, 250K - 950K parameters) embeddings models called BERT Hash https://huggingface.co/blog&#
7.
by dmezzetti 3mo ago
Why does HN let OpenAI and Anthropic basically advertise but it throws down the gauntlet at a small developer like myself when we do "self promotion"? Top 3 posts as of this moment are all about Claude.
8.
by dmezzetti 3mo ago
Local models are great for a lot of things past just software development. We need to move towards solving other real world problems vs just building software. I've been focused on that with TxtAI ( https://github.com/ne
9.
by dmezzetti 3mo ago
Well-Read Students Learn Better: On the Importance of Pre-training Compact Models Related paper that's a good read: https://arxiv.org/abs/1908.08962
10.
by dmezzetti 7mo ago
They have paid hosting - https://huggingface.co/enterprise and paid accounts. Also consulting services. Seems like a pretty good foundation to me.
11.
by dmezzetti 7mo ago
This is really great news. I've been one of the strongest supporters of local AI dedicating thousands of hours towards building a framework to enable it. I'm looking forward to seeing what comes of it!
12.
by dmezzetti 7mo ago
Why keep relying on API services? If you'd like your own local AI integration with open providers like Rocket.Chat and Mattermost, check out txtchat ( https://github.com/neuml/txtchat ).
13.
by dmezzetti 7mo ago
Very interesting! It would be great to see how it compares to Faiss / HNSWLib etc. I'd will consider integrating it into txtai as an ANN backend.
14.
by dmezzetti 8mo ago
Have you considered using something like claude code / opencode?
15.
by dmezzetti 8mo ago
It's been an issue for a while and it's even bigger now in the age of AI. Lots of people use security as a way to "have their moment" and don't really care about adding value. But scaring people off from security re
16.
by dmezzetti 8mo ago
It's too bad that Anthropic is so hostile to open source. It's a big missed opportunity for them.
17.
by dmezzetti 8mo ago
Nice application, great work!
18.
by dmezzetti 8mo ago
AI Development is good for those who want to do it. But not a terminal career decision for those who don't.
19.
by dmezzetti 8mo ago
Two words: Open Source.
20.
by dmezzetti 9mo ago
This change was happening well before LLMs. People were tired of being yelled at and treated poorly. A cautionary tale for many of these types of tech platforms, this one included.
21.
by dmezzetti 9mo ago
The message here is good. I've now spent over 5 years in the OSS world ( https://github.com/neuml ). I started by picking a problem I was interested in and checking the work into GitHub. I've been extremely fortunat
22.
by dmezzetti 9mo ago
I love this feature. I've long used json_extract to create dynamic columns with txtai sql: https://neuml.github.io/txtai/embeddings/query/#dynamic-colu... You can do the same with DuckDB and Postgres too
23.
by dmezzetti 10mo ago
I've used SQLite as the content storage engine for years with TxtAI. It works great. Also plenty of good add-ons for it such as sqlite-vec for storing vectors. It can take you pretty far and maybe it's all you need in many circums
24.
by dmezzetti 10mo ago
Looking forward to trying them out. Great to see they are Apache 2.0...always good to have easy-to-understand licensing.
25.
by dmezzetti 10mo ago
Companies that consider an open source project a critical part of their infrastructure should sponsor or compensate those projects. Also when someone finds a bug, the maintainers are under no obligation to fix it or fix it with any timeline
26.
by dmezzetti 10mo ago
Are multiple LLM queries faster than vector search? Even with the example "dog OR canine" that leads to two LLM inference calls vs one. LLM inference is also more expensive than vector search. In general RAG != Vector Search thoug
27.
by dmezzetti 10mo ago
I agree. I did this myself with TxtAI. It can store vectors, data, graphs and keyword indexes all to Postgres. https://medium.com/neuml/postgres-is-all-you-need-for-vector...
28.
by dmezzetti 10mo ago
Fun project. I'm sure it will get a lot of interest here. For those into vector storage in general, one thing that has interested me lately is the idea of storing vectors as GGUF files and bring the familiar llama.cpp style quants to i
29.
by dmezzetti 10mo ago
Glad to see all the interest in the local RAG space, it's been something I've been pushing for a while. I just put this example together today: https://gist.github.com/davidmezzetti/d2854ed82f2d0665ec7efd...
30.
by dmezzetti 10mo ago
I've said it multiple times. I don't want to force those who use my projects to have to share their code unless they want to. But if someone wants to use GPL that's up to them. It's a choice.

More ›