Can Agentic AI be used to augment existing tools? I think there is a lot of nice open source software out there that is too complex/difficult/heavy for most users. There is an opportunity to use AI to simplify the user experience through some automation.
This is the little experiment I did here, building an AI-SAST by wrapping CodeQL with an agentic layer. It makes it so easy to use CodeQL that I’m now revisiting past projects and discovering quite a few surprises!
The agent is used to:
- write custom codeql queries based on the code and a prompt
- triage SARIF output and group issues, extract most important ones
- investigate issues for risk, false positive, etc
I built a simple tool to help software development teams benchmark their AI adoption (in a AI Native Dev spirit).
Every team says they’re “doing AI,” but maturity levels vary a lot. This tool asks 12 short questions across 6 dimensions (coding, process, governance, people, culture, value). It then gives a 1–5 maturity score, plus a radar chart showing strengths and gaps.
I originally built it for conversations with companies, VCs and dev teams I work with — but I thought it might be useful for others here as well.
I’d love feedback:
- Do the questions feel relevant to real-world teams?
- What dimensions do you think are missing?
Thanks for taking a look!
PS: Share your scores below and let's discuss how to improve :-)
Thanks, I just tried it out to generate some content for my blog. I like the flow and the quality of the output which sounds much better than usual AI generated BS.
One thing that didn't work are the references. All errored (Information not found).
One thing I missed is the ability to add some custom data. In this case I would have like to give a few points on what my product is, what keywords are important for SEO, what is the spin I'd like to give the post.
In Back to the Future, what was Doc Brown's dog's name in 1955? A seemingly simple movie trivia question exposes fundamental limitations in current RAG systems. Answering this proves surprisingly challenging for our current technology.
The fact that I don't find any flaw in the reasoning shows one thing: either this new o1 is crazy good and this getting pretty close to AGI... or I'm too bad at maths
Hi HN, I'd love to get your thoughts on this one! Anyone using LLM, hidden inside an app, just as a reasoning 'brick' to progress some workflows, decide on best math, etc.
Could you detail what you mean by deploying LLMs ? Is it about integrating commercial LLMs in an enterprise context? Or running self-hosted LLM for a small company (e.g. Ollama + Ollama Web UI)? Or integrating an Agentic approach to existing software stack?
However, the issue you will quickly encounter is resources/costs. For a simple mode like llama3-7b you need at least a g5.2xlarge on AWS. If you want a 'chat gpt equivalent' model, you need something like llama3-70b or command-r-plus etc. These will require at least a g5.48xlarge that will cost you $20 an hour
An alternate approach is going hybrid: self-hosted UI which takes care of user access, shared documents (RAG), custom prompts etc but that you hook to a LLM provider where you pay per token (could be OpenAI platform or anything from Huggings).
Let me know if this helps! Also note that I'm lead dev of an open source project addressing these kind of needs: https://opengpa.org - Feel free to jump on our Discord to discuss.
If you are already on AWS, then it's better to run llama3 using AWS Bedrook. With Bedrock, you only pay for what you use instead of paying for an always-on EC2 instance.
I've written a short fiction setting the scene for my open-source project and how it could fit in a corporate/enterprise context. If you are in such an environment, I'd love to hear from you and better understand your needs and concerns in implementing your company AI strategy!
Hi HN! I'm the creator of OpenGPA, an open-source General Purpose Agent platform.
Key features:
- Self-hostable ChatGPT-like agent
- Extensible with custom Actions for enterprise data/APIs
- Built in Java/Spring for easy enterprise integration
I built this to explore agentic AI development and its potential in enterprise software. Looking for feedback, use cases, and contributors.
What challenges do you see in integrating AI agents into enterprise systems? What features would be most valuable to you?
This is the little experiment I did here, building an AI-SAST by wrapping CodeQL with an agentic layer. It makes it so easy to use CodeQL that I’m now revisiting past projects and discovering quite a few surprises!
The agent is used to:
- write custom codeql queries based on the code and a prompt
- triage SARIF output and group issues, extract most important ones
- investigate issues for risk, false positive, etc
- prepare pull requests with fixes
Github: https://github.com/eschnou/patchsmith
Blog post with more details: https://blog.transcode.be/patchsmith-ai-codeql-cli/
What other interesting open source tool could benefit from a AI driven UI/UX wrapper layer?