yosai

Cool. Hydron can index PDFs up to 5,000 pages. Roughly how long does it take to index a 5000 page PDF?

Thanks for validating the pain point..I was also using chatGPT.It's was a pain.We are building something really big to solve the problem of the system developers.Stay tuned.

Devin and Cognition are great as co-workers for application software. They're proficient in JavaScript and Python and excel at solving logical problems in software. However, they aren't trained in or familiar with system software, which typically runs on physical devices and requires verification against technical specifications for programming.

Hey cjtechie,It's not an another co-pilot.Existing coding co-pilot can only do code generation,code comprehension But not allow you to upload your technical datasheet or crash log etc.Most of the coding agents are SaaS based and SOC2, GDPR and even HIPAA certified. However, these certifications do not guarantee data privacy since they are relatively easy to obtain. Moreover, most SaaS make use of third party providers for different sub-processes (LLMs, embeddings, reranking…), resulting in private data being transferred and stored on numerous servers across the internet.privacy-aware coding SaaS services anonymize sensitive data before sending it to AI. This approach combines the ease of adoption and power of SaaS with data privacy. However, it comes with a single point of failure, which is the anonymization algorithm itself.We solve it by genarting coding datsets and finetuning a small open source coding model on-prem.You can compare the response in a side by side view to gpt-4o model in our platform.

Yes rnavi,I have the same experience as well.The complete system software development can be expedited using AI.Yes, we are doing private beta with a Japanese semiconductor conglomerate.We go with a philoshopy like your data, your model and your cloud.We create instruction dataset from their tech spec and code.Everything on-prem.We have partnered with AMD and Nvidia for on-prem GPU deployment.We help them selecting the right small language coding model and deploying it with securely.

We use multimodal RAG and tools similar to unstructued.io ,We generate structured output and use LLM again to do the matching with our AST parsed source code.Now matching part is really complex and need manual inspection and validation.

Thanks shrianshag..You can try some sample Datasheets(pdf) and ask code generation questions.

We have a team of domain expert who do the vetting of the instruction dataset.We do typical RLHF(Reinforcement learning from human feedback) and connect back to our SFT(supervised finetuning) loop.That's why we name ourself as hardware and human in loop.Humans play an important role in ensuring quality and accuracy of our dataset.

Looks to be a great observability tool.Any specific reason to choose rust here?