A GPT to determine if a product should be considered Ultra Processed food or not. Give it the URL to a web page with nutritional information, or just input the ingredients, or even a picture of the ingredients from the label.
I've made some modifications to this GPT and seems to work pretty well. It takes a list of ingredients you provide via text, URL or image and gives a response which includes whether the product is "ultra-processed" and what it's NOVA classification is.
It also provides a brief explanation of why it made that determination.
I'm not sure what I think about this, but here we are.
TL;DR: Explore Legalpioneer Copilot, a new resource with 600 law-related GPTs covering legal, tax, and regulatory issues. Find and access GPTs created by individuals and companies for legal information and support.
I tried to help a friend develop a #GPTs that could query the outside world through #web#APIs.
It wouldn’t just do what you wanted. You couldn’t just provide it with <fill the blanks> url.
Declaring the call had to be made with intense difficulty. Every parameter declared with precise description formally, so it could guess the string right.
Then it occurred to me: They didn’t want you to succeed, they wanted you as #free#labour to #train their #AI first, only second to do your business.
The demise of chatGPT may be in progress. I get a variety of errors when I try to access it right now. And I've had a few over the weekend. It was working as expected earlier today.
Are other chatGPT plus users having similar issues?
In the future, how much is the popularity of a programming language, framework, or library affected by how well LLM's like ChatGPT or GitHub Copilot know how to use them? Having good documentation and many examples openly available is beneficial, so the LLM’s can learn the tool. Does this benefit established tools over new ones?
@kellogh I've been testing the #GPTs feature: it currently doesn't work with brief documentation, hallucinating incorrect ways to use a library. It seems to work better with clear examples (as many as possible) and descriptions instead of just technical function specifications, but these are time-consuming to do.
If the underlying model has learned "a common way" to do something, then a library doing it very differently might be in trouble, even if the model is given proper instructions. 🤷
@mikkohei13@kellogh ..i am experimenting too, i found the #GPTs i was trying to make performed better giving it instructions to look to specific URLs/webcontent instead of adding public documents with the same content to it's "knowledge". GPT4 itself (not GPTs) works excellent when adding a document in a query , but i can only do that with public documents for obvious reasons.. 🤔
Many claim the GPT Store is an App Store Moment, but not so fast. I dive into this claim, along with how it’s easier to steal and monetize other’s content with GPTs, and if you have IP or confidential information, you should certainly avoid them. Security still isn’t a consideration, much less an afterthought. Quite a few will learn this lesson the hard way. #GPTs#GPTStore
The space of custom GPTs seems to be really exploding, curious what the effect will be, not only on OpenAI financial results but also on society. Good example, a curated list of GPTs for #cybersecurity. On the darknet there will be such lists for malware and bad purposes i can imagine... 🤔 https://github.com/fr0gger/Awesome-GPT-Agents #AI#ChatGPT#GPTs
Here is what it had to say about #IOCX :
“IOC.exchange is a platform designed for the sharing of Indicators of Compromise (IOCs) among cybersecurity professionals and organizations. It's a collaborative environment where participants can contribute, exchange, and access IOCs related to various cybersecurity threats. These IOCs typically include data such as malicious IP addresses, URLs, file hashes, and email addresses associated with cyber threats.
The platform's primary goal is to aid in the early detection and response to cyber threats by providing a repository of actionable intelligence. By aggregating and sharing IOCs, cybersecurity teams can more effectively identify and mitigate potential threats to their networks and systems. It's important to note that the effectiveness of such a platform largely depends on the quality, relevance, and timeliness of the shared information.
[…]”
Is anyone else who is experimenting with openai's new #GPTs experiencing a lot of network errors and other inconsistencies? I am hoping that this is a temporary thing as the service becomes more robust.
Now using paid ChatGPT aka ChatGPT plus as an experiment, quite impressed with GPT4 not so much yet with "GPTs" , but maybe i am not good enough in instructions etcetera. #AI#ChatGPT#GPT4#GPTs
I wonder if universities could create a custom GPT for each course to provide students with a tutor. Supply the GPT with all course materials, related readings, problem sets, and instructions, and let students use it as a 24/7 available TA.
I quickly created a new Assistant with the OpenAI Playground for my Meta Minutes video podcast. I gave it some detailed instructions and provided the episodes of all 4 seasons with titles, guests, questions and descriptions. 🤖
Not only can I find things faster now with semantic search but it can also help me to get ideas for new episodes. Love it. 😍
OpenAI is making the platform play, ofcourse you can use it's API's but it clearly wants people to develop "GPTs" on it's own platform and sell them in the store. It will be interesting to see whether this will fly. Or that we just get swamped by useless , hallucinating stochastic parrots 😀 #OpenAI#AI#GPTs