Mar 2022 - doraithodla/techlog GitHub Wiki

Mar 2022

  • https://www.affectiva.com/

  • Everything Old Is New Again - The Blog of Matt https://bit.ly/3tn5gmr

  • Enjoy programming https://bit.ly/3JDCO5v

  • Building games and apps entirely through natural language using OpenAI’s code-davinci model – @AndrewMayne https://bit.ly/3ikMvcI

  • http://semantictwitter.appspot.com/

  • PythonC/Python Cookbook, 3rd Edition_ Recipes for Mastering Python 3 ( PDFDrive ).pdf at main · aibooks14/PythonC https://bit.ly/3u6YCzE

  • How to Connect Python to Google Sheets | Coupler.io Blog https://bit.ly/37CCPIG

  • Pipedrive Dashboard To Track Progress | Coupler.io Blog https://bit.ly/3N1CpvM

  • Microsoft Power Platform Center of Excellence Kit - Power Platform | Microsoft Docs https://bit.ly/3tbAdtr

  • https://www.visualthesaurus.com/

  • Programmable Blocks Tease Self-Assembling Space Structures - IEEE Spectrum https://bit.ly/3KIJGyw

  • https://github.com/tweag/nickel/?utm\_source=tldrnewsletter

  • Token Engineering

  • How to make and use argument maps : Avoid endless discussions - YouTube https://bit.ly/36cD6Bs

  • https://oli.cmu.edu/courses/visual-design-open-free/

  • https://en.wikipedia.org/wiki/Argument\_map

  • Improving Data Quality with Product Similarity Search | by Evi Lazaridou | commercetools tech https://bit.ly/3MLU8Hp

  • Lots of things are turning up during RA and Happi research. Need to organize them for reuse

  • Reading about software engineering for Exeter

  • Going to get Soumya focused on developing front ends till her engineeing skills improve.

  • Need to collect a set of rules for refactoring python code. There is an entire book on refactoring javascript

  • Need to collect a list of Python idioms (alredy have a book on Python patterns)

  • todo: build a mind map generator from text

  • The concept of multiple stages of conversation is pretty cool. It applies to many types of (chatbot) conversations. I am thinking of a few:

    • Customer conversations (sales, support)
    • Tutor conversations (where you are teaching some one something)
    • Career conversations
    • product conversations

  • "Using pretrained language models is the fastest way to perform common NLP tasks. In contrast, if you need to perform uncommon NLP tasks, you may need to train the model from scratch, including sourcing and annotating/labeling the data relevant for your task." #AdvancedNLP

  • It matters to talk about the 99% Developers because these are the developers building the software that powers our lives — insurance, health care, retail, and banking, just to name a few. It’s not only small companies that can’t easily adopt the processes of modern, tech-first companies; it’s most companies that were not built around technology and that have decades of legacy software practices firmly in place. Many of these companies move around quite a bit of money. Many of these companies handle quite a bit of our personal data. If technology innovations are not benefiting these software teams, we’re losing out on a lot of meaningful improvements to everyone’s quality of life. https://future.a16z.com/software-development-building-for-99-developers/

  • 99% Developers

  • Dark Matter Developers: The Unseen 99% by Scott Hanselman

  • Code coverage Google blog post - https://bit.ly/3pIqXer

  • a16z blogs on Data Archives | Future https://bit.ly/3tEf4XS

  • Book : Probabilistic Models of Cognition - 2nd Edition https://bit.ly/3sTJUMZ

    • Computational theory of mind
    • Generative models

  • It is generally preferable to create new virtual environments for every machine learning project you have. Having a separate environment for each project allows you to install the relevant libraries for your current project without having to uninstall libraries that you may need for other projects but that can cause code to fail for your current project. Think of a virtual environment as a blank canvas (i.e., new set of libraries) for you to do your work without having to worry about how changes to the current canvas conflict with canvases for your other projects.

  • Transfer learning is a way of taking a pre-trained model and customizing it for a specific need. This is what we will be doing with our various GPT-3 products.