The Daily Index — Monday, January 6, 2025 · TLDR Index

Across today’s newsletters

Today’s Index

TLDR

TESLA CYBERTRUCK SALES ARE DISASTROUS

Cybertruck sales are stalling a year into production. While Tesla claimed to have over 1 million reservations for the vehicle, it had lowered its reservation deposit to just $100 and the production version of the truck was more expensive and had less range than what Tesla originally announced. The company delivered around 40,000 Cybertrucks before opening orders beyond the reservation program. It's estimated that Tesla delivered between 9,000 and 12,000 Cybertrucks in Q4.

electrek.co · 3 min

TLDR AI

XAI'S GROK MODEL FACES DELAY

xAI has delayed the launch of its next-gen Grok model, citing quality concerns, marking yet another delay in the AI industry.

techcrunch.com · 4 min

TLDR Web Dev

BUILDING A POLITE & FAST WEB CRAWLER

A Mozilla engineer noticed that poorly-behaved bots, primarily from OpenAI and Amazon, caused 70% of the load on Diapsora's servers (an open-source social network). To address this, he built a “polite and fast” web crawler using Python and gevent, prioritizing respectful rate limiting by defaulting to one request per domain every five seconds and dynamically adjusting based on server responses. The crawler efficiently manages the queue by fetching robots.txt files in parallel and uses a sophisticated refetching strategy using HTTP headers to minimize redundant requests.

cameronboehmer.com · 4 min

TLDR InfoSec

TIME TO CHECK IF YOU RAN ANY OF THESE 33 MALICIOUS CHROME EXTENSIONS

Researchers discovered 33 compromised Chrome extensions affecting 2.6 million devices during the holiday season. The incident began when Cyberhaven's extension was hijacked through a phishing attack on Christmas Eve, leading to malicious code that stole sensitive data including Facebook and ChatGPT credentials.

arstechnica.com · 6 min

TLDR Crypto

$19 TRILLION IN TRANSACTIONS SETTLED ON THE BITCOIN NETWORK IN 2024

Bitcoin's network activity doubled in 2024 as it processed $19 trillion in transactions compared to 2023's $8.7 trillion. The network also achieved a record hashrate of 1,000 EH/s in early 2025, with US-based mining pools controlling 40% of the global hashrate. Chinese pools maintain majority control.

cointelegraph.com · 2 min

TLDR

SAM ALTMAN ON CHATGPT'S FIRST TWO YEARS, ELON MUSK, AND AI UNDER TRUMP

OpenAI launched ChatGPT on November 30, 2022. Within two months, its site was being visited by more than 100 million visitors. The startup hasn't been the same since. This article contains an interview with OpenAI's CEO Sam Altman. Altman explains his infamous four-day firing, how he runs OpenAI, his plans for the Trump-Musk presidency, and his relentless pursuit of artificial general intelligence.

bloomberg.com · 39 min

TLDR

THIS YEAR, RISC-V LAPTOPS REALLY ARRIVE

Framework plans to release a laptop that can support a RISC-V mainboard in 2025. RISC-V is an open-source instruction set architecture (a set of valid instructions programs can execute on a processor) that provides a viable alternative to x86 and Arm. It is open source, making it easy to customize hardware. While current RISC-V chips perform well behind x86 and Arm-powered alternatives, performance will be much better in the coming year.

spectrum.ieee.org · 4 min

TLDR

SPACEX ANNOUNCES SLATE OF UPGRADES FOR ITS NEXT STARSHIP LAUNCH

SpaceX has made special improvements for its seventh Starship test flight, which may be on January 10. The test will attempt Starship's first payload deployment test, multiple reentry experiments geared towards ship catch and reuse, and the launch and return of the Super Heavy booster. The booster will splashdown in the Gulf of Mexico if the catching systems aren't healthy during the launch. The booster will slow down from supersonic speeds upon return, which will result in sonic booms around the catch area.

links.tldrnewsletter.com · 3 min

TLDR

CAN LLMS WRITE BETTER CODE IF YOU KEEP ASKING THEM TO “WRITE BETTER CODE”?

There was a trend a while ago where people kept asking AI image generators to make something 'more X', which interestingly resulted in the images eventually becoming something cosmic. This article looks at whether this concept could apply to large language model-generated code - whether asking an LLM to 'make code better' can result in improvements. It appears that asking an LLM to 'write code better' results in better code compared to direct prompting (at least with Claude 3.5 Sonnet). LLMs can produce interesting ideas and tool suggestions even if the output can't be used as is.

minimaxir.com · 22 min