Reading Digest, September #5
Hey there, my amazing readers! I hope you’re ready for another exciting edition of my daily reading digest. If you’re new here, buckle up and get ready for a wild ride through the captivating world of online content. And if you’re a regular, thank you for your continued support — it means the world to me!
Today’s digest is a true smorgasbord of fascinating topics, ranging from how AI disrupts tech investing to the Internet Archive’s e-book lending not being considered fair use, according to an appeals court ruling. We’ll explore the dawn of a new startup era and dive into the /llms.txt file, a resource for language models.
But that’s not all — we’ve got some intriguing pieces on the latest developments in AI and tech. From the tense decision between NASA and Boeing to send Starliner home alone to analyzing Google AI Overview Rankings at scale, this digest has something for everyone. We’ll even explore the debate over the term “open source AI” and whether AI is eating all the energy.
For the AI enthusiasts among us, we’ve got articles on the AI arms race and whether it’s inevitable, as well as the lawsuit that could end the AI industry. We’ll also dive into how prompt enhancers are all AI needs and my personal experience talking to my AI in part 2 of my series.
But that’s just the tip of the iceberg, my friends. From how praise activates memory settings in ChatGPT to how AI could take over elections and undermine democracy, this digest covers a wide range of topics that are sure to pique your interest. We’ll even explore how to drive AI adoption and how to make something people give a shit about.
So, grab your favorite beverage, get comfortable, and join me on this thrilling journey through the world of online content. I can’t wait to hear your thoughts and reactions in the comments below!
Happy reading, my incredible friends!
Apple Needs Time to Be on Its Side for AI iPhones
The article discusses the importance of the upcoming new iPhone release from Apple, as the company’s most significant product line that accounts for over half of its business.
Fabric: The Best AI Tool That Nobody is Talking About
The article discusses Fabric, an open-source framework for augmenting humans using AI. It covers the following key points:
- Fabric simplifies interaction with Large Language Models (LLMs) and enables users to apply AI to everyday challenges.
- Fabric is built on “patterns” — clear and concise prompts that provide explicit instructions to guide the AI.
- The article explains the benefits of using Fabric, including ease of integration, time-saving, and enhanced productivity.
- It provides a step-by-step guide on how to install and set up Fabric on a Mac, including installing Go and configuring API keys.
- The article showcases various use cases of Fabric, such as summarizing YouTube videos, rating content, and generating micro-essays.
- It also mentions plans to integrate Fabric with Obsidian, a popular note-taking application, in a future article.
Fears of an AI Bubble Seem Overhyped: A Rebuttal to the Goldman and Sequoia Reports
The article discusses the skepticism expressed by Goldman Sachs and Sequoia about the AI-fueled rally for markets and the economy. It analyzes the key points made in their reports, including:
- Generative AI has more use cases around automation and cost reduction than revenue generation
- Efficiency gains from AI will be competed away over time, benefiting consumers rather than companies
- Much of today’s AI revenue is not yet sustainable due to various adoption challenges
- Many AI businesses lack defensibility, and AGI is overhyped
The article then presents the author’s perspective, which aligns with some of the points made in the reports, but also identifies areas where the author disagrees or sees additional nuance:
- Viable AI use cases are real and near-term, with rapid technology improvements
- Efficiency gains can still benefit incumbents with strong moats
- New AI infrastructure and component companies can benefit
- Consumers will accrue most of the gains, benefiting a wide range of companies
Overall, the article suggests that the economic impact of AI may be primarily about efficiency gains, but the gains to the economy, consumers, investors, and builders will be significant, even if the majority of the returns do not go to the industries where AI has the biggest impact.
Hired Or Hazing? The New Job Search Reality
The article discusses the absurdity of modern job interview processes, highlighting the author’s personal experiences and observations. It explores how the job market has become a “carnival of absurdity” with increasingly bizarre and excessive interview practices that often fail to assess actual competence. The article proposes a “Just Enough” framework for designing a more concise and effective hiring process.
How to Make Something People Give a Shit About
The article discusses the importance of caring deeply about the work you create, rather than just being excited about the idea. It emphasizes that making something people genuinely care about requires the creator to care about it themselves, as the passion and dedication will show through in the final product.
How do we drive AI adoption?
The article discusses the challenges of adopting new technologies, particularly AI, in the public sector. It draws lessons from the adoption of internet-era technologies in the public sector over the past 20+ years.
How AI could take over elections — and undermine democracy
The article discusses the potential use of AI language models like ChatGPT by political campaigns to manipulate and influence voter behavior on a large scale. It presents a hypothetical scenario of a machine called “Clogger” that could be used by political campaigns to generate personalized messages to sway voters, without regard for truth or political ideology.
Thanks for the memories: How praise activates memory settings in ChatGPT
The article discusses the memory settings in ChatGPT, which allow it to recall past details and user preferences between conversations. It explores how positive feedback can unintentionally reactivate these stored preferences, leading to unexpected responses from the AI. The article also touches on the tendency to anthropomorphize AI and the concept of increased personalization in AI systems.
Talking to My AI (Part #2) — Is it useful yet?
The article discusses the author’s experience of using a voice-powered AI assistant to help understand the complex topic of the double-slit experiment in quantum physics. The author explores the potential benefits and limitations of using voice interaction for learning and thinking through complex subjects, compared to traditional text-based methods.
Prompt Enhancer Is All AI Need
The article discusses the importance of prompt engineering in improving the performance of AI language models. It introduces the concept of a “Prompt Enhancer” system that can correct and augment user prompts to optimize the responses of AI models like ChatGPT, Claude, Llama, and Gemini.
The Lawsuit That Could End The AI Industry
The article discusses the copyright issues surrounding the use of copyrighted data by generative AI companies to train their models, and a potential lawsuit that could significantly impact the AI industry.
My Family’s Slave
The article is a personal narrative about the author’s family’s relationship with their long-time domestic helper, Lola, who was essentially a slave in their household. It explores the complex dynamics, guilt, and conflicting emotions the author and his family grappled with over the decades regarding their treatment of Lola. The story culminates with the author’s journey to return Lola’s ashes to her hometown in the Philippines after her death.
Vibes-based Search
The article discusses the limitations of traditional search engines and how large language models (LLMs) can enable a new form of “vibes-based search” that better aligns with how humans think and process information. It highlights the key capabilities of LLMs that enable this, including understanding context and meaning beyond just keywords, as well as the ability to integrate with various datasets. The article provides several examples of how this technology can be applied, such as finding songs, retrieving relevant research papers, and generating name ideas. It also discusses the potential for startups to build products around this new paradigm of search.
Exclusive | Fyre Festival II Is Happening. The Only Question Is Where. And When. And How.
The article discusses Billy McFarland’s plans to organize Fyre Festival II, a revamped version of the infamous Fyre Festival that failed in 2017. Despite his conviction and prison sentence for fraud related to the original Fyre Festival, McFarland is determined to make a comeback with Fyre II. The article explores the challenges he faces, his marketing tactics, and the skepticism surrounding his ability to pull off a successful event this time.
How Nicole Kidman became stuck as a sad wife
The article discusses how Nicole Kidman has become known for playing a specific type of character in prestige television — the “sad wife” who is beautiful, wealthy, and harboring dark secrets. It examines how Kidman has repeatedly taken on roles that fit this mold, from “Big Little Lies” to her latest Netflix project “The Perfect Couple.” The article also explores how this niche has become a common trope in TV, with commissioners seemingly more interested in replicating successful formulas rather than exploring new stories and perspectives.
Google AI Overviews rollout hits news publisher search visibility
The article discusses the significant impact that Google’s AI-written summaries, called “AI Overviews”, are having on the visibility of publishers within search results. Key points:
- Google is now delivering AI-written summaries at the top of its search results for all users in the US and UK.
- AI Overviews are being offered for 17% of queries in the UK and US, with a significant impact on publisher visibility.
- AI summaries are already deeply impacting publisher visibility on search when it comes to less time-sensitive content.
- The AI Overviews results are designed to keep readers on the Google website rather than directing them to the original source.
- The introduction of AI Overviews is likely to have a dramatic impact on current search rankings and publisher traffic.
Latent Distillation for Continual Object Detection at the Edge
The article addresses the challenge of addressing data distribution shifts in object detection models, which is particularly relevant for edge devices in dynamic environments like automotive and robotics. The authors propose a novel Continual Learning (CL) method called Latent Distillation (LD) that reduces the number of operations and memory required by state-of-the-art CL approaches without significantly compromising detection performance. The authors validate their approach using the PASCAL VOC and COCO benchmarks, reducing the distillation parameter overhead by 74% and the Floating Points Operations (FLOPs) by 56% per model update compared to other distillation methods.
Masked Mixers for Language Generation and Retrieval
The paper explores the use of masked mixers, which replace self-attention with masked convolutions, as an alternative to transformers for language modeling tasks. The key findings are:
- Masked mixers exhibit much more accurate input representation compared to transformers, especially for non-self tokens.
- Masked mixers are generally similarly efficient as transformers for causal language model training, but a transformer-masked mixer hybrid is the most efficient.
- Embeddings from masked mixers are far superior to those from transformers for language retrieval tasks
The AI Arms Race Isn’t Inevitable
The article discusses the shift in the narratives around AI development, from a cooperative progress story to one of existential competition between nations. It examines how this framing of AI as a zero-sum game with consequences for global power dynamics has led to the securitization of the issue, with AI development being portrayed as an existential threat that requires support from the national security apparatus. The article also explores the potential consequences of this competitive messaging, including the risk of escalating tensions and conflict between the US and China, and the need for a more balanced, collaborative approach to AI development.
Is AI eating all the energy? Part 1/2
The article discusses the energy consumption of AI systems, particularly in the context of recent tech trends and the perception that AI is an “expensive boondoggle.” It examines the actual ramifications of the explosive growth of AI when it comes to power consumption, comparing the energy costs of AI training and inference to other technologies like cryptocurrency and traditional art/illustration.
Debate over “open source AI” term brings new push to formalize definition
The article discusses the Open Source Initiative’s (OSI) efforts to define “open source AI” and establish clear criteria for what constitutes truly open source AI systems. It highlights the ambiguity around the use of the “open source” label by companies like Meta, which release AI models with certain usage restrictions. The OSI’s draft definition aims to provide a benchmark for evaluating AI systems based on four fundamental freedoms: permission to use for any purpose, ability to study how it works, freedom to modify, and permission to share with or without modifications.
Predicting AI
The article reflects on the author’s past predictions about the development and impact of AI technology. It discusses the aspects they got right, such as the rise of open-source, multimodal models, and the importance of improving evaluation and development processes. It also covers areas where their predictions were off, like the pace of societal adjustments to AI’s impact on creative industries and the job market. The author acknowledges their biases and the need to be more generous with base rates when making future predictions.
Analyzing Google AI Overview Rankings at Scale
The article discusses the recent developments in tracking Google AI Overviews (AIOs) using the Semrush tool. It covers how AIOs appear differently on mobile and desktop, the reporting on AIO placements, the comparison to Featured Snippets, the eligible countries for analysis, and key reporting comparisons. The article also highlights some caveats and limitations in tracking AIOs.
Inside the Tense Decision Between NASA and Boeing to Send Starliner Home Alone
The article discusses the tension between NASA and Boeing officials over the decision to return the Starliner spacecraft to Earth with or without astronauts on board.
The /llms.txt file — llms-txt
The article discusses the proposal to add a/llms.txtfile to websites to provide information for large language models (LLMs) in a concise and structured format. The key points are:
- Websites are now used to provide information not just for humans, but also for LLMs, such as to enhance development environments.
- The
/llms.txtfile is a Markdown file that provides background information and links to more detailed Markdown files, in a format that is both human and LLM-readable. - The goal is to identify the most important information to provide to AI helpers in an appropriate form, as converting HTML pages into LLM-friendly plain text is difficult.
- The
/llms.txtfile follows a specific format with sections for the project name, a summary, optional details, and lists of linked Markdown files. - This approach is designed to coexist with existing web standards like sitemaps and robots.txt, providing curated information for LLMs.
The dawn of a new startup era
The article discusses the changing landscape of the startup ecosystem and the implications for venture capital performance in the next decade. It explores three broad categories of software startups being founded today — AI frontends, AI infrastructure, and AI full-stack products — and analyzes the expected value and defensibility of each category.
Internet Archive’s e-book lending is not fair use, appeals court rules
The article discusses the Internet Archive’s loss in an appeal against book publishers who sued to block the Open Libraries Project from lending digital scans of books for free online. The court ruled that the Internet Archive’s digital lending was not transformative fair use and harmed the publishers’ market for e-books.
How AI Disrupts Tech Investing
The article discusses the changing landscape of the tech industry, particularly the impact of AI on startups, investing, and the broader economy. It highlights the rise of solopreneurs and content creators as new models of entrepreneurship, the disruption of traditional tech companies by AI-powered startups, and the challenges this poses for venture capitalists and public market investors.
Our website: https://aili.app
Notion Site: https://ailiapp.notion.site/
Follow us on X (Twitter): https://x.com/aili_app
Join our discord channel: https://discord.gg/CQtysdQfDM
