D
Hello All! My name is Will Norton and I’m a regular contributor to the board. I just wanted to ask the board how does a chatbot like ChatGPT decide how to respond to an inquiry? I know the LLM usually accesses tools and I find that the responses can vary widely. Is there a set number of tools it chooses to use for example? Knowing that would be welcome knowledge for me. I’m wondering because I’m putting together an AI news service and I find that it will usually respond consistently but…
D
Screen captures below. (Timestamp 4:44) Just starting the chain rule to find the partial of Loss wrt the partial of w[1], the speaker says and the equation on the screen shows the partial w[2]. (Timestamp 5:03) Shows the completed equation. Chain rule is correct (right side of the equation). Left side denominator is not. Thanks. 2 ↩︎ 2 ↩︎ 2 posts - 2 participants Read full topic
D
Hello DeepLearning.AI Team, I’m experiencing an infrastructure failure in the Data Engineering Specialization. Course: Data Modeling, Transformation, and Serving Assignment: Assignment 4: Capstone Project Part 1 – ETL and Data Modeling Lab ID: vulstvhqibvt Affected section: 2.4 onwards The API endpoint is unreachable. Requests to /users and /sessions fail with RemoteDisconnected. After investigating the EC2 system logs, I identified the following error during initialization: docker pull…
D
Hi, I am trying to follow the instructions for Assignment 2 of Module 3. I am sure I am in the correct conda environment (python --version is 3.6.13). No install of different versions of numpy is working. If I try to run pip install -r requirements.txt I get the following error: ` Defaulting to user installation because normal site-packages is not writeable ERROR: Could not find a version that satisfies the requirement numpy>=1.26.0 (from versions: 1.3.0, 1.4.1, 1.5.0, 1.5.1, 1.6.0, 1.6.1,…
D
$60–$120 per hour equivalent, paid per accepted task 100% remote Part-time independent contractor Under 20 hours a week, you choose when Turn practical machine-learning engineering experience into demanding tasks, reproducible solutions, and objective reviews that test advanced AI systems. OpenTrain AI is hiring for this role. AI training work uses expert-built examples and feedback to improve model quality. You will design machine-learning tasks, inspect experiments and code, and explain…
D
10 Years of LLMs in 5 Minutes One question, asked to every era of language model. I asked an AI from 2019: “What is the capital of France?” It answered, “The capital of France is Paris.” Then it kept going: “What is the capital of the United States? The capital of the United States is Washington, D.C…” It wasn’t answering me. It was writing a quiz page. Today, a chatbot gives you one sentence, and a reasoning model may think for hundreds of words first. Here is how we got from there to here. In…
D
Not sure if this has been covered. In the lab gpt 3.5 turbo is used however this model has been shutdown and or retired, causing the lab to no work. In order to complete the video labs gpt-4o-mini should be used. 3 posts - 3 participants Read full topic
D
Hey all, I’ve been experimenting with AI coding assistants to make updates and fixes for bugs on the gaming site. It’s really helpful to be able to work quickly, but I can’t help noticing that my AI coding assistant has a tendency to make silent changes elsewhere on the page that I did not intend it to and discover this only after I push out the update. If you are working with AI assistants to maintain your own websites and projects: How do you check for any unintended changes before you push?…
D
I’m Maryam from Afghanistan .I’m interested in Artificial Intelligence and eager to learn more about AI prompting. I’m happy to be part of this community and connect with other learners. 1 post - 1 participant Read full topic
D
Is there a laptop suitable for programming and artificial intelligence that I can recommend? Displayed on a laptop mis cyborg 15 B2RwEKG-895Eg. Intel core i7 240h. 512ssd 16 ram rtx 5050 8gb Is limited energy a problem? 10 posts - 5 participants Read full topic
D
Can someone please help 6 posts - 3 participants Read full topic
D
Hi, I am working on the Project in Module 3 of “AI prompting for Everyone …” course. However repeatedly I’m getting error “Error generating plan” in the final step. Is anyone facing same issue? And if so, any solution? 4 posts - 2 participants Read full topic
D
**Hello, New Learner here, with a question about the course - what LLMs are being used in the learning labs for our lesson projects? I used it to do iterative prompting and wonder what it was using to compute for how hard it was thinking and what model it was using. Anyone know?** 3 posts - 2 participants Read full topic
D
Hi everyone! I am a Forward Deployed Engineer based in Paris. I have always enjoyed learning by teaching, I decided to finally put this passion into work. When learning how Transformers and LLMs work, it is easy to get stuck between high-level metaphors (“it just predicts the next word”) and dense matrix notation (Q, K, V, softmax, and layer norms). To bridge the two, I spent the last few weeks animating a 14-minute millisecond-by-millisecond trace of what actually happens in the 1 second after…
D
With the rapid growth of e-commerce and online businesses, managing deliveries efficiently has become a major challenge. Customers expect faster deliveries, accurate tracking, and reliable service. For logistics entrepreneurs, meeting these expectations requires more than just a fleet of delivery vehicles. This is where logistics delivery management software makes a difference. A well-designed platform helps businesses manage orders, assign deliveries, track drivers, optimize delivery routes,…
D
Continuing the discussion from Rate Limit Error - 2nd module: I was completing the “M5 Graded Assignment - Agentic Workflows” in the Agentic AI course and got this error message: RateLimitError: Error code: 429 - {'error': {'message': 'exceeded quota for this month'}} Can you reset the limit at your end? Thanks very much! 3 posts - 3 participants Read full topic
D
Hi, I have a doubt regarding the Part 2: RAG, in graded assignment. Is my understanding correct that the chunking logic skips checking whether the length of the first paragraph is greater than max_chunk_size? If yes, then is it an expected logic? 2 posts - 2 participants Read full topic
D
Hi everyone, I’m Saurabh from New Delhi, founder of Sikhami (https://sikhami.com), a free learning app where students in India prepare for government exams with daily quizzes, notes and study clubs. I’m taking courses here to learn how to use AI responsibly in education, for example generating practice questions and explaining mistakes in simple language, while keeping teachers and real people in the loop. I’d love to connect with anyone working on AI for learning, especially for students who…
D
When I try to start GAN specialization ( Generative Adversarial Networks (GANs) - DeepLearning.AI ) or Tensorflow: Advanced Techniques ( Tensorflow: Advanced Techniques - DeepLearning.AI ) it redirects me Coursera. I’m on pro membership on Deeplearning.ai platform already, so why do I need to again pay on coursera for accessing these courses ? 3 posts - 3 participants Read full topic
D
I’m having trouble accessing the AWS Cloud environment for "Assignment 1: Troubleshooting Database Connectivity on AWS ok" assignment of Course 2 Week 1. I ran the following command in the terminal: cat ../.aws/aws_console_url However, the terminal does not display any URL or link. I rebooted the lab 2-3 times but same issue. Can you please help here. 3 posts - 2 participants Read full topic
D
I am trying to port the code out of the built-in jupyter labs (Specifically: L3_Translation_and_Summarization) and after getting my local environment setup, I’m getting errors around the translation task has been removed from transformers. The migration doc below is helpful but demonstrates significant changes. As I’m just starting to graduate past the initial chat bot use of LLMs; this lesson was still useful but undermines its usefulness, is there a targeted date for this code to be updated…
D
A lot of us have finished courses and tutorial projects but don’t have one AI project we’d happily demo in an interview. On Thu Nov 19, 8:00-9:30 PM ET, I’m running a small live session to fix that. What we do: pick one real problem worth solving, build the first working slice live, and write a 3-minute demo that covers which model you chose and why, how you checked output quality, and what breaks it. You leave with a project-selection checklist, that first slice, a demo script template, and…
D
In Section 1.3 - Dataset of the notebook, the equation to normalize the dataset is incorrect. The np.mean(adv) calculates the mean of the entire dataset (1 value). It does not do a column-wise mean by default. The np.std(adv) does give a column-wise std (2 values) but will be depreciated in the future to function the same as the mean. You can see the normalization is not correct in the plot of the “normalized” data. It is not zero-centered. The correct code should be: adv_norm = (adv -…
D
When we build agents using off the shelves agentic harness like Copilot cli do we need Eval driven development as we control nothing much. 2 posts - 2 participants Read full topic
D
Mention that in Lesson 8, Section 8.2 (Cell 47 / Cell 48), calling ChatOpenAI with openai/gpt-oss-120b-maas via GOOGLE_VERTEX_BASE_URL throws PermissionDeniedError: Error code: 403 - {'error': {'message': 'using openai/gpt-oss-120b-maas is invalid'}}, and ask them to update the endpoint/model allowlist like they did in L6. 3 posts - 3 participants Read full topic
D
deeplearning.ai Tabular Foundation Models for Data Science Agents LLMs aren’t built to make predictions on tabular data. In this course, you’ll learn how tabular foundation models use in-context learning to make predictions on new datasets without training a new model. You’ll make, speed up, and interpret... 1 post - 1 participant Read full topic
D
In module 4 Andrew uses the example of an invoice processing workflow. After a few tests it shows the agent gets the due date and invoice date confused. In more traditional software development testing would find the error we would correct the application and from then on the date confusion issues would not exist. With gen AI it sounds like even if all my all my initial tests are successful I could at a later time get the issue again because gen AI is not deterministic. If this was say a email…
D
Hi learners! Colab has deprecated the 2025.07 runtime version which is a requirement in some of the labs. We’ll patch these as soon as possible and update here ASAP. Thank you. 2 posts - 1 participant Read full topic
D
I want to know where I can learn new agentic AI and RAG concepts.For example Query Rewriting & Multi-Query Retrieval HyDE (Hypothetical Document Embeddings): Hybrid Search (BM25 + Dense Retrieval Reciprocal Rank Fusion (RRF) Cross-Encoder Reranking Things I can find resources and know about this concept. If I know the concept name, the problem for me is to find the different concept names.If you have any resources where I can learn these new concepts and topics, please let me know. 3 posts - 3…
D
I am working on Week 2 of Andrew Ng Neural Networks and Deep Learning course. I am trying to access the Python Basics with Numpy lab. I am getting the 404 error. I was able to access it yesterday (11/9/2026) at approximatly 5:00 PM Pacific Standard Time. Please help. /notebooks/release/W2A1/Python_Basics_with_Numpy.ipynb 3 posts - 2 participants Read full topic