sudoapps

6 ms·

1.
by sudoapps 5mo ago
Awesome! Let me know if you run into any issues with the setup.
2.
by sudoapps 5mo ago
Coding agents have changed how I build. Constantly switching between the terminal and an IDE started to feel inefficient, so I wanted a better terminal-first setup where I could manage multiple agent sessions and make quick edits without th
3.
by sudoapps 5mo ago
Coding agents should help us reduce dependencies overall. I agree Go is already best positioned as a language for this. Using random dependencies for some small feature seems archaic now.
4.
Show HN: Lightweight agent proxy for monitoring sensitive data exposure(github.com)
1 points by sudoapps 6mo ago|0 comments
5.
by sudoapps 7mo ago
Hot take - Local LLM computing will move to stationary, always on devices (Mac mini & studio). Developers and users will move to lighter, portable devices to interface with their long running agent workers (MacBook Airs & iPads).
6.
Codex built my startup in a weekend(sudoapps.substack.com)
3 points by sudoapps 7mo ago|1 comments
7.
by sudoapps 7mo ago
Codex built my old startup in a weekend. What took me a few months to build 2+ years ago can now be done with coding agents in 2 days. The last couple months feel like a step change with Codex and Claude.
8.
Show HN: Night Watch, zero-dependency DevOps agent(github.com)
2 points by sudoapps 7mo ago|0 comments
9.
Show HN: Small Hours, An AI-powered observability and root cause analysis(smallhours.dev)
3 points by sudoapps 2y ago|0 comments
10.
by sudoapps 3y ago
How do others feel about the pace of innovation in AI? Is there ever a reason to slow down advancements in a given field?
11.
by sudoapps 3y ago
Speed and performance improvements like this are why VSCode has passed Atom and other editors over time.
12.
Gorilla LLM: Going deeper with the team at UC Berkeley(sudoapps.substack.com)
2 points by sudoapps 3y ago|0 comments
13.
Gorilla LLM: An interview with the team at UC Berkeley(sudoapps.substack.com)
2 points by sudoapps 3y ago|0 comments
14.
by sudoapps 3y ago
Very similar to a consensus network. That is probably the future of LLM agents.
15.
by sudoapps 3y ago
This was a quick experiment on creating multiple LLM (GPT) agents with their own objective to see if outputs can be improved.
16.
The Age of Agents: Can multiple LLM agents improve outputs?(sudoapps.substack.com)
2 points by sudoapps 3y ago|3 comments
17.
by sudoapps 3y ago
I can see this being useful as a specialized (fine-tuned) LLM in a chain of LLMs for full autonomy
18.
The Age of Agents(sudoapps.substack.com)
2 points by sudoapps 3y ago|1 comments
19.
by sudoapps 3y ago
Autonomous agents could be the next evolution of AI, extending the capabilities of LLMs. Interested to see which agent implementations seem the most promising today.
20.
by sudoapps 3y ago
As this LLM agent architecture continues to evolve and improve, we will probably see a lot of incredible products built on top of it.
21.
by sudoapps 3y ago
This is really interesting. Could this lead to eventually being able to deconstruct these "black-boxes" to remove proprietary data or enforce legal issues?
22.
by sudoapps 3y ago
Completely agree
23.
by sudoapps 3y ago
This wasn't mean't to say that all training would stop. I think, to some extent, the model won't need additional recent data (that is already similar in structure to what it has) to better understand language and interpret th
24.
by sudoapps 3y ago
OpenAI doesn't let you fine-tune GPT-4 or GPT-3.5 yet ( https://platform.openai.com/docs/guides/fine-tuning ), but fine-tuning models on a set of documents is still an option but not really scalable if you want
25.
by sudoapps 3y ago
> But then I see model context length getting longer and longer just within the transformer architecture and the training engineering going on. Do you have any references to this? Seems really interesting if that can be a long term appro
26.
by sudoapps 3y ago
The article is definitely still high level and mean't to provide enough understanding of what capabilities are today. Some of what you are mentioning goes deeper on how you take these learnings/tools and come up with the any numbe
27.
by sudoapps 3y ago
A lot of what prompting has turned into seems silly to me too, but it has shown to be effective (at least with GPT-4).
28.
by sudoapps 3y ago
Agreed, GPT answering based on its own training data has been the best experience by far (aside from hallucinations) and comparing against that is difficult. Embeddings might not even be the long term solution. I think it's still early
29.
Giving GPT “Infinite” Knowledge(sudoapps.substack.com)
121 points by sudoapps 3y ago|86 comments
30.
by sudoapps 3y ago
If you are wondering what the latest is on giving LLM's access to large amounts of data, I think this article is a good start. Seems like this is a space where there will be a ton of innovation so interested to learn what else is comin

More ›