AI and LLMs

Rooms with Text: A Dataset for Overlaying Text Detection

In this paper, we introduce a new dataset of room interior pictures with overlaying and scene text, totalling to 4836 annotated images in 25 product categories. We provide details on the collection and annotation process of our dataset, and analyze its statistics. Furthermore, we propose a baseline method for overlaying text detection, that leverages the character region-aware text detection framework to guide the classification model. We validate our approach and show its efficiency in terms of binary classification metrics, reaching the final performance of 0.95 F1 score, with false positive and false negative rates of 0.02 and 0.06 correspondingly.

Why Are There So Many Rationalist Cults?

There’s a lot to like about the Rationalist community, but they do have a certain tendency to spawn — shall we say — high demand groups. We sent a card-carrying Rat to investigate what’s really going on.

The Global Brain Argument: Nodes, Computroniums and the Ai Megasystem (Target Paper for Special Issue)

The Global Brain Argument contends that many of us are, or will be, part of a global brain network that includes both biological and artificial intelligences (AIs), such as generative AIs ...

Reify This

The authors contend that contemporary efforts to render AI systems interpretable rest on a mistake: reification, the process of treating abstractions and statistical artifacts as if they were concrete realities.…

"Restoration is possible, but not by nostalgia" – Luciano Floridi on

In conversation with Infomedia, the University of Bergen's newest honorary doctor discusses the erosion of the public information space, responsible AI, and what education should look like in an age

"Restoration is possible, but not by nostalgia" – Luciano Floridi on

In conversation with Infomedia, the University of Bergen's newest honorary doctor discusses the erosion of the public information space, responsible AI, and what education should look like in an age

We read a lot of AI books so you don’t have to

“Because every camp needs a book right now”

Slop-Machine Future

The arc of large language models is mediocre, and it bends toward “target procurement”.

Solving a Million-Step LLM Task with Zero Errors

LLMs have achieved remarkable breakthroughs in reasoning, insights, and tool use, but chaining these abilities into extended processes at the scale of those routinely executed by humans,...

A Dark-Money Campaign Is Paying Influencers to Frame Chinese AI as a Threat

Build American AI, a nonprofit linked to a super PAC bankrolled by executives at OpenAI and Andreessen Horowitz, is funding a campaign to spread pro-AI messaging and stoke fears about China.

A red pixel in the snow: How AI solved the mystery of a missing mountaineer

Scouring remote areas for missing walkers and climbers can take rescuers weeks and sometimes months. AI can do the job in a matter of hours in some cases – and potentially save lives.

AI-slop , GrantaGate and Bad Writing

AI writes poorly because it has crammed too much fanfic jargon

What Can We Learn From the FDA Model for AI Regulation? - AI Now Institute

After holding a rapid deliberation late last year that convened deep experts on the FDA alongside key participants in the AI policy debate, we pulled together several immediate insights into a memo. Though it was clear the path forward isn’t to port over any single regulatory model wholesale, hosting a deep dive into the Food […]

The Case That A.I. Is Thinking

ChatGPT does not have an inner life. Yet it seems to know what it’s talking about.

Pretext Does What CSS Can't — Measuring Text Before the DOM Even Exists | HackerNoon

Cheng Lou's Pretext library measures multiline text height without touching the DOM — unlocking layout capabilities CSS has never been able to offer.

Universal AI Bypass: How Policy Puppetry Leaks System Prompts and Safety Data

HiddenLayer’s latest research uncovers a universal prompt injection bypass impacting GPT-4, Claude, Gemini, and more, exposing major LLM security gaps.

How LLMs are and are not like the brain

Hi from buttondown! At the bottom of this newsletter is a bit of administrivia about the new platform How LLMs are and are not like the brain Beneath all the...

How StrongDM’s AI team build serious software without even looking at the code

Last week I hinted at a demo I had seen from a team implementing what Dan Shapiro called the Dark Factory level of AI adoption, where no human even looks …

StrongDM Software Factory

StrongDM's field notes on non-interactive agentic development: specs + scenarios, validation harnesses, feedback loops, and the supporting components.

Mercor’s 23-Year-Old Billionaire Founders Grapple With Employee Fraud And North Korean Infiltration

Founded in 2023 by 20-somethings, data labeling startup Mercor exploded to $1 billion in annualized revenue run rate earlier this year. Now it's confronting a wave of challenges, including an employee stealing money, security blunders and cultural growing pains.