🚀 Early Access! Many things may still not work as I refactor the site and make improvements. - Learn more

2025-04-15

A Talmudic Rationalist Cautionary Tale

Published on April 15, 2025 4:11 AM GMTThe following is a translated famous passage from the Babylonian Talmud (Bava Metzia 59a–b) that I believe has a good implicit rationalist intepretation....

3 clicks (3 unique) 1 month ago

Creating 'Making God': a Feature Documentary on risks from AGI

Published on April 15, 2025 2:56 AM GMTPlease donate to our Manifund (as of 14.04.25 we have two more days of donation matching up to $10,000). Email me at connor.axiotes...

2 clicks (2 unique) 1 month ago

MindEye2 (Not Pretrained) vs. MindEye1

In this section, we show how MindEye2 outperforms MindEye1 even without pretraining on other subjects.

5 clicks (5 unique) 1 month ago

Today's Wordle answer for Tuesday, April 15

Get all the help you need with today's Wordle.

4 clicks (4 unique) 1 month ago

A Dissent on Honesty

Published on April 15, 2025 2:43 AM GMTContextDisney's Tangled (2010) is a great movie. Spoilers if you haven't seen it.The heroine, having been kidnapped at birth and raised in a...

5 clicks (5 unique) 1 month ago

MindEye2: Shared-Subject Models Enable fMRI-To-Image With 1 Hour of Data: Single-Subject Evaluations

In this section, we show more exhaustive evaluation metrics computed for every subject individually using 40 hours and 1 hour of fine-tuning data, respectively.

3 clicks (3 unique) 1 month ago

How AI Agents Could Have Supercharged Pfizer’s COVID-19 Vaccine Development

AI agents are systems that can reason, act, and learn over time. They aren’t just passive tools. They’re active collaborators. We use them in software engineering to automate code reviews,...

5 clicks (5 unique) 1 month ago

Reconstruction Evaluations Across Varying Amounts of Training Data: Mindeye2

Here, we present a further analysis of how model performance scales with training data.

3 clicks (3 unique) 1 month ago

Principles for coding securely with LLMs

Writing code with LLMs is fundamentally different from other ways of programming. LLMs are often non-deterministic and always unpredictable. They have a capability that no other technology can match: the...

5 clicks (5 unique) 1 month ago

ChatGPT Can Now Remember and Reference All Your Previous Chats

Samuel Axon, writing for Ars Technica: Some time ago, OpenAI added a feature called “Memory” that allowed a limited number of pieces of information to be retained and used for...

5 clicks (5 unique) 1 month ago

Young men are 'playing videogames all day' instead of getting jobs because they can mooch off of free healthcare, claims congressman

To justify Medicaid cuts, House Speaker Mike Johnson says the healthcare program leads to "able-bodied young men" playing videogames instead of working.

4 clicks (4 unique) 1 month ago

MindEye2 Dataset: fMRI Preprocessing and Normalization

Provides further details about the fMRI dataset used in MindEye2, including preprocessing and normalization procedures.

3 clicks (3 unique) 1 month ago

Behind the Scenes: The Making of MindEye2

Details the contributions of each author involved in the development and research of MindEye2.

3 clicks (3 unique) 1 month ago

Becoming a Writer in a Hyperconnected World

All platforms are now engaging, interactive. We can collaborate everywhere. To know the best platforms for evolving, we need to reflect on how we interact with these systems. Platforms, like...

4 clicks (4 unique) 1 month ago

Correcting Deceptive Alignment using a Deontological Approach

Published on April 14, 2025 10:07 PM GMTDeceptive alignment (also known as alignment faking) occurs when an AI system that is not genuinely aligned behaves as if it is—intentionally deceiving...

4 clicks (4 unique) 1 month ago

If Getting Laid-off is Bad, Wait Till You See What AI Does to Corporates

AI advancements present a bigger threat to bureaucracy corporates than it does to highly-skilled professionals.

3 clicks (3 unique) 1 month ago

What if there was a nuke in Manhattan and why that could be a good thing

Published on April 15, 2025 12:19 AM GMTIf a country wanted to perform a nuclear first strike, launching a missile might be the wrong way to do it.What if instead,...

4 clicks (4 unique) 1 month ago

Developer Productivity with GitHub Copilot & AI Tools by Aditya Mishra

As a runner-up in R Systems BlogBook, Aditya Mishra unpacks GitHub Copilot’s impact on coding workflows—from AI-powered suggestions to code quality, testing, and learning.

3 clicks (3 unique) 1 month ago

🛠️ Fixing the Decline: 5 Proven Ways to Boost Credit Card Authorization Rates

Credit card authorization failures quietly destroy revenue across the internet. Most of them aren't due to fraud. And most can be prevented. Let's walk through five strategies modern merchants use...

5 clicks (5 unique) 1 month ago

Balatro's publisher beat gold stakes and the Jokerless challenge so quickly it even surprised LocalThunk: 'How did you do this so fast?'

"I sent him the video of the score, going, like, screw you. Yeah, I got it!" said Playstack's Wout van Halderen.

5 clicks (5 unique) 1 month ago

Welcome to Postreads

Discover and follow the best content from across the web, all in one place. Create an account to start building your personalized feed today and never miss out on great reads.

Support Postreads

Enjoying the service? Help me keep it running and improve it further by buying me a coffee!

Buy me a coffee

Content Timeline

\