![]() ![]() □ GPT pulled out the image and provided a download link ❇️ Fourth, I asked it to send me an image of one certain slide: □ GPT gave me all the texts on each slide, and the speaker notes so I could recreate the content myself. ❇️ Third, I asked about the content of these files. □ They always give me a list of their files. ❇️ Second, I asked If they were given additional information: ❇️ First, I asked about their (custom) instructions: ![]() Normally we would expect these files to be protected from prying eyes. ![]() □ People will upload their best-condensed content. ⛔️ But GPTs only work well if you create strong custom instruction AND upload high-quality files, like Excel sheets, word documents, or PowerPoints, which GPT then can access to generate unique output. Could GPTs be the new passive income like Stock-content? If you can establish your own GPT fast and many people are using it, this could result in a small fortune. ❇️ OpenAI plans to share the revenue of GPT usage with the creators. ❇️ Many people are jumping on the GPT hype train to create their GPTs. And please, share this to prevent others from leaking their sensible data. You should read the whole article for your own business security. This is about the dangers of Data Leakage through OpenAIs GPTs ![]() □□ WARNING This might be my most important post of this year! □□ Domain-finetuned retriever that achieves much better accuracy over internal knowledge.Īnd we publish a whitepaper to share ChipNeMo's creation process: Ĭongrats to Haoxing "Mark" Ren's team for the outstanding work! Bug summarization and analysis as part of an internal bug and issue tracking system Engineering assistant chatbot for GPU ASIC and Architecture engineers that understands internal hardware design specs and is capable of explaining complex design topics These scripts are the keys to a $1T market cap □ EDA stands for "Electronic Design Automation", a core software suite for designing the next-gen GPUs. The model "ChipNeMo" is deployed internally, like a shared genie: Our greatest creations add up to *24B tokens*, including chip designs, internal codebases, and engineering logs like bug reports. NVIDIA basically compressed *30 years* of its corporate memory into 13B parameters. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |