The Daily Index — Monday, August 21, 2023 · TLDR Index

Across today’s newsletters

Today’s Index

TLDR AI

3 TRILLION OPEN TOKEN DATASET RELEASED

There are many open datasets, but few of sufficient scale to train a frontier model. The Allen Institute for AI’s Dolma dataset aims to enable researchers to study data effects at scale.

blog.allenai.org · 4 min

TLDR Web Dev

A GUIDE TO OPTIMIZING JAVASCRIPT FILES

Optimizing JavaScript files resolves the issue of render blocking, page load time, and other common web performance issues. This article delves into practical JavaScript file optimization techniques, such as tree shaking, and goes through tools to help the optimization process. It also provides tips to help debug performance issues.

sitepoint.com · 10 min

TLDR Cybersecurity

MICROSOFT: SPHYNX RANSOMWARE EMBEDS IMPACKET, REMCOM

Microsoft's Threat Intel team posted a series of blogs about BlackCat’s new version of its Sphynx encryptor. Among the changes in the new version: popular post-exploitation tools Impakcet and RemCom were bundled with the encryptor.

bleepingcomputer.com · 4 min

TLDR Crypto

DAY

Coinbase's offshore derivatives exchange has seen a surge in daily trading volumes, reaching up to $287m. Launched in May, the platform has onboarded 50 institutions and generated $5.5B in volume in its first seven weeks. The aim of this product is to accelerate Coinbase's global expansion amidst challenging regulatory conditions domestically.

theblock.co · 2 min

TLDR Founders

1998

China is looking to attract foreign investment after its recovery from Covid times has been slower than hoped.

reuters.com · 1 min

TLDR AI

OPENAI’S CONTENT MODERATION

OpenAI proposes using GPT-4 for content policy development and content moderation decisions across digital platforms, taking the burden away from humans and allowing more consistent labeling of content and faster feedback loops. This begs the question, should AI be the authority on who can say what?

openai.com · 5 min

TLDR AI

SIGMOID LOSS FOR LANGUAGE IMAGE PRETRAINING

CLIP was somewhat revolutionary for computer vision. It has some extremely clever algorithmic tricks (e.g., large in-batch negatives) that make it perform well. This paper questions many of those tricks and finds a simpler training setup that outperforms CLIP for the same task.

arxiv.org · 30 min

TLDR AI

AN EFFICIENT WAY TO EVALUATE TEXT-TO-IMAGE GENERATIONS

Current methods for judging the quality of images created from text have some problems, like not truly capturing how good the image looks or how well it matches the text. This paper introduces a new method that estimates how likely a created image matches a given text, focusing on the most important parts of the image. It does this efficiently with fewer samples.

arxiv.org · 17 min

TLDR AI

IMPROVING 3D DETECTION IN NEW ENVIRONMENTS

Detecting objects in 3D using LiDAR has challenges when moving to new places because of differences in how data looks. This new method, GPA-3D, uses special reference points to understand 3D shapes better and helps the detector to work well even in unfamiliar areas.

arxiv.org · 16 min