LLM Transformer Model Explainer:
Nice tool for understanding Transformer-based AI models.
06.03.2025 13:19
👍 2
🔁 0
💬 0
📌 0
LLM Transformer Model Explainer:
Nice tool for understanding Transformer-based AI models.
5 ChatGPT Hacks That You’ve Never Heard of (But Need to Know)
More useful tips, including some I didn't know.
4 ChatGPT Features You May Not Know About
Stuff you may not have even noticed! Max your outputs:
What Is Retrieval-Augmented Generation aka RAG?
Really useful explanation of how to improve AI accuracy using added knowledge sources
Generative AI for Beginners:
21-lesson program on Github for learning AI development using practical projects.
World's Largest Call Center Deploys AI to 'Neutralize' Indian Accents
This is an example of AI stuff that gives me the creeps.
AI Agent collection:
A dedicated network for AI agents and autonomous systems you can rent.
agent.ai
A Curated List of AI Agent Resources ->
Nice Github collection of research papers, projects, frameworks and tools:
buff.ly/HU1LOyh buff.ly/fhVKsce
Finally to close this thought provoking special issue, @irisvanrooij.bsky.social and coauthors present "Reclaiming AI as a theoretical tool for cognitive science" where the authors stand firmly against the hype and harms of current AI practice. I was the proud guest editor of this incisive article.
Make sure to look into the company beforehand. As someone who conducts interviews, it really bugs us when candidates come to an interview without any knowledge of the company, or without any questions about the job, culture, or anything. Here's how to research a company
My own thinking on this is far from complete. Still, my intuition tells me that refusing to guide students on ways AI can complement their learning and thinking is ultimately a dead end. I've tried to candidly articulate my reasons in a draft directed at students: docs.google.com/document/d/1...
We have a properly warped perception of how much various sectors contribute to the UK economy. The fishing industry, which is constantly in the news and to whom politicians of all rosettes regularly play, adds less to GDP than Games Workshop
1/2 Valuable piece. TL;DR summary in chart.
Through past year plus, MAGA messaging was nonstop "worst economy in history," and press coverage revolved around "Yeah, this really feels terrible for everyone. [Have you seen the price of eggs?]"
To mention positive trends was to be Marie Antoinette.
Both GPT-4o and Claude were unreasonably good at helping me write the whole quiz.
Core skills:
* Suggesting people, places, topics, events from computing history
* Thematic round ideas like the anagrams round, visual picture round, and 'guess the language' round
* Clever puns for round names
Yes :-) You can see this very thing with tax changes. Rather than pick mechanisms (a) easily understood by the public (b) supported by the public and (c) bonehead easy to implement and without risk of foment - nah, let's pick some new territory and do a bit of cheese paring on all of them 😂
This is neat - I gave Claude the famous bombers-with-red-dots image and the prompt "create a simulation as an artifact that will illustrate the point of this image." Then I said "make it better"
Now this is a much more intuitive explanation than a meme! Play with it here claude.site/artifacts/ca...
This paper on Claude with Computer Use matches my experience- as a general purpose agent that can do anything in a computer, it is surprisingly good. Yet it still has enough flaws that it is a sign of the future than a full agent now.
But it also shows the future is soon. arxiv.org/pdf/2411.103...
True. After running workshops for nearly 900 people, at least 80% of those who were sceptical, turned out to be those who had simply never tried (much) to use these tools.
AI can help learning... when it isn't a crutch.
There are now multiple controlled experiments showing that students who use AI to get answers to problems hurts learning (even though they think they are learning), but that students who use well-promoted LLMs as a tutor perform better on tests.
Regardless of your opinion on how well they are doing, I'm old enough to spot exactly the same ambitious but doomed technocrats as I spoke to in '96 & '97. All fired up to manage society, people, problems and annoyances away with the theatre of useless tweaking - rather than real structural work.
Book outline
Over the past decade, embeddings — numerical representations of machine learning features used as input to deep learning models — have become a foundational data structure in industrial machine learning systems. TF-IDF, PCA, and one-hot encoding have always been key tools in machine learning systems as ways to compress and make sense of large amounts of textual data. However, traditional approaches were limited in the amount of context they could reason about with increasing amounts of data. As the volume, velocity, and variety of data captured by modern applications has exploded, creating approaches specifically tailored to scale has become increasingly important. Google’s Word2Vec paper made an important step in moving from simple statistical representations to semantic meaning of words. The subsequent rise of the Transformer architecture and transfer learning, as well as the latest surge in generative methods has enabled the growth of embeddings as a foundational machine learning data structure. This survey paper aims to provide a deep dive into what embeddings are, their history, and usage patterns in industry.
Cover image
Just realized BlueSky allows sharing valuable stuff cause it doesn't punish links. 🤩
Let's start with "What are embeddings" by @vickiboykis.com
The book is a great summary of embeddings, from history to modern approaches.
The best part: it's free.
Link: vickiboykis.com/what_are_emb...
Fewer than 20 tickets remaining!
thirteen.barcamplondon.org
We're putting on an unconference and want *you* to come along.
An unconference is an event where *you* set the agenda. Come talk about whatever you're interested in. Learn about what other people are passionate about.
Win LEGO!
#BCLxiii
5 Devices to Rule Them All: This is the Testing List you are Looking for -> www.linkedin.com/pulse/5-devi...
generative-models.jpg
1/ X's algorithm was changed in mid-July 2024 to systematically boost Republican-leaning accounts and Elon Musk's own account following his endorsement of Donald Trump, according to a newly released computational study of engagement from the Queensland University of Technology.⬇️
