Topic
Gpt
7 episodes
-
Overview: Transformer Architecture
We finally sit down and define transformer architecture from the ground up, because we keep throwing the term around like it’s obvious and it really isn’t. We use the attention-as-a-room-of-index-cards picture to make the mechanism click, then connect it to why Transformers became the backbone of modern language models.
-
Overview: Tokenization
We slow down and explain tokenization from the ground up: how raw text becomes numbered pieces a model can process, why those pieces are usually subwords, and why the tokenizer quietly affects cost, context, language handling, and product behavior.
-
Beyond Alignment: Value Diversity as a Collective Property in Multicultural Agent Systems
A research paper from Singapore University of Technology and Design and Washington University in St. Louis introduces 'value diversity' as a system-level evaluation metric for multicultural multi-agent systems. The core finding: existing LLM-based agent systems are systematically less diverse than human societies and show almost no correlation between per-agent cultural alignment and system-wide value heterogeneity. Single-backbone systems fall far short of human diversity levels (36.12 vs. 44.07); mixed-backbone configurations help but don't close the gap; and social interaction between agents drives homogenization rather than preserving plurality. The paper uses the World Values Survey across 19 cultures and 18 models, includes a participatory budgeting case study, and releases code and datasets.
-
ARIS: Autonomous Research via Adversarial Multi Agent Collaboration
Justy and Cody dig into ARIS, an open-source harness for autonomous ML research that assumes a single long-running agent will eventually make unsupported claims. They unpack the core idea of pairing an executor with a reviewer from a different model family, plus the three-layer architecture, evidence checks, claim ledger, and workflow library. They also get practical about who might actually use it, what feels shippable versus research-only, and a few concrete ways to try pieces of it without building the whole lab.
-
DV World: Benchmarking Data Visualization Agents in Real World Scenarios
Justy and Cody dig into DV-World, a new benchmark from a multi-institution research team that stress-tests AI data visualization agents on real-world tasks — spreadsheet manipulation, cross-framework chart evolution, and handling ambiguous user intent. Even the best models top out around 50%, which tells you a lot about where the gap actually is.
-
SketchVLM: Vision language models can annotate images to explain thoughts and guide users
In this episode, Justy and Cody dig into SketchVLM, a training-free framework that lets vision-language models explain answers by drawing editable SVG annotations on top of images. They talk through why text-only answers are hard to verify, how SketchVLM uses a draft-and-refine loop plus visual grounding to produce overlays, where it looks production-friendly, and where the trade-offs still show up.
-
The Three Ages of Data Science: When to Use Traditional Machine Learning, Deep Learning, or an LLM (Explained with One Example) | Towards Data Science
Data Science The Three Ages of Data Science: When to Use Traditional Machine Learning, Deep Learning, or a LLM (Explained with One Example) A practical use case to describe how the data scientist job changed across three generations of machine learning Piero Paialunga Nov 11, 2025 10 min read Share Photo by Markus Spiske on Unsplash One of the best songs of the universe (made by one of the most iconic singers ever) says this: Wish I could go back And change these years I’m going through changes Black sabbath – Changes This song is incredibly powerful and talks about how life can change right in front of you so quickly. That song is about a broken heart and a love story.