The Daily Index — Wednesday, May 17, 2023 · TLDR Index

Across today’s newsletters

Today’s Index

TLDR

IMPLOSION IS EVEN BETTER

FTX's founder, Sam Bankman-Fried, is the focus of several charges leveled by federal prosecutors. These include money laundering, bribery, and orchestrating a fraudulent scheme that resulted in the multibillion-dollar misappropriation of customer money. Bankman-Fried has pleaded not guilty. Michael Lewis was originally planning to just check out Bankman-Fried as a favor to his friend in the fall of 2021, but he ended up receiving a front-row seat to the scandal. His book about Bankman-Fried is set to be released on October 3.

archive.ph · 6 min

TLDR AI

TINY STORIES, 1 TRANSFORMER BLOCK CAN GENERATE STORIES

With models getting larger and more expensive we generally get better performance. But what happens if we go the other way? It turns out that tiny models can learn to generate coherent, novel, and grammatically correct stories. If you carefully construct the training data set and restrict the vocabulary to common words, you get some compelling completions from these miniscule models.

arxiv.org · 25 min

TLDR Web Dev

MY 20 YEAR CAREER IS TECHNICAL DEBT OR DEPRECATED

With enough time, most code becomes technical debt or deprecated. The author, over his 20 year career, reflects on how many of the lines of code, projects, and languages he’s worked on have either been deprecated or rewritten. This is not necessarily a bad thing but the reality of being a software engineer.

blog.visionarycto.com · 4 min

TLDR Crypto

IMPLOSION IS EVEN BETTER.

As was reported last year, Michael Lewis is releasing a book about the rise and fall of Sam Bankman-Fried, the founder of FTX, one of the largest cryptocurrency exchanges. Titled "Going Infinite: The Rise and Fall of a New Tycoon," the book recounts how Lewis, initially introduced to Bankman-Fried by a friend, became an observer of an unfolding scandal. FTX faced a spectacular collapse, with Bankman-Fried accused of money laundering, bribery, and fraudulent schemes. Lewis, having shadowed Bankman-Fried for about a year, captured the unfolding events. The book is due to release on October 3.

archive.ph · 6 min

TLDR Founders

WHICH AI MODEL SHOULD YOU PICK FOR YOUR STARTUP?

Tomasz Tunguz helps answer a question that many PMs face today: to use a small language model or a large language model? In the realm of AI, the decision between using a certain model is influenced by factors such as time constraints, reliance on external expertise, infrastructure management, ip, data sensitivity, and cost, while the market is still evolving, and managed large-language models currently have an advantage in terms of adoption and revenue.

tomtunguz.com · 2 min

TLDR

THE RACE TO BRING GENERATIVE AI TO MOBILE DEVICES

Generative AI may be soon integrated into mobile. The push towards advanced AI in software and services has resulted in increased computing costs for tech companies. Operating generative AI on mobile devices could resolve one of the main economical issues with the technology. Most of the work on tailoring models to handsets is still at an experimental stage, so it is too early to see whether these efforts will lead to truly useful mobile applications.

archive.ph · 7 min

TLDR AI

IMPROVING ACTION RECOGNITION: EXPLORING MULTIMODAL GENERALIZATION

This study introduces a new concept in egocentric action recognition called Multimodal Generalization (MMG), which focuses on how systems can adapt when data from certain sources is limited or missing. To investigate this, a new dataset called MMG-Ego4D was created, and new methods were developed to enhance the system's ability to generalize, potentially guiding future research in this area.

arxiv.org · 9 min

TLDR AI

SUPERICL: MAKING BIG AI MODELS WORK BETTER WITH SMALLER ONES

SuperICL is a new technique that helps massive AI models, like GPT-3 and GPT-4, to learn better by working with smaller, locally fine-tuned models. This approach not only boosts the performance of these big AI models on tasks, but also improves their stability and the abilities of smaller models, such as understanding different languages and explaining their decisions.

arxiv.org · 6 min

TLDR AI

THE FIRST RNN ON THE TRANSFORMERS LIBRARY

The RWKV model is a linear attention model with built in recurrence. It is a powerful language model that runs fast with a theoretically long context window. The model development has an active discord community which now is more visible on the HuggingFace platform.

huggingface.co · 15 min