The hottest Research Substack posts right now

And their main takeaways
Category
Top Health Politics Topics

Roche Nanopore: First Thoughts (Pt 1)

ASeq Newsletter โ€ข 51 implied HN points โ€ข 20 Feb 25
๐Ÿ”ฌ Science Genetics Biotechnology Research Innovation Technology
  1. Roche Nanopore's new sequencing method shows really impressive early results, which builds confidence in its technology.
  2. The accuracy of their duplex sequencing is reportedly better than Illumina's, suggesting it could be useful for many different applications.
  3. More information will be shared soon, as there's ongoing discussion and a second post in the works.

Barcoding Brains

Asimov Press โ€ข 225 implied HN points โ€ข 09 Feb 25
๐Ÿ”ฌ Science Neuroscience Biotechnology Health Research Technology
  1. Connectomics is a method that maps connections in the brain, helping scientists understand how these connections affect our behavior and brain function. However, it's really slow and expensive to do.
  2. A non-profit called E11 Bio is developing new tools to make brain mapping faster and cheaper. They aim to create a complete map of a mouse's brain in just five years for a fraction of the current cost.
  3. Understanding the brain's connections can improve treatments for various brain diseases and potentially lead to advanced technologies like brain-computer interfaces. This could really change how we approach brain health and artificial intelligence.

Building a phosphorescence detector

lcamtufโ€™s thing โ€ข 2652 implied HN points โ€ข 27 Nov 24
๐Ÿ”ฌ Science Physics Engineering Research Technology Materials
  1. There isn't much information available on natural phosphorescence, so creating a detector can help explore its presence in everyday items.
  2. The device design prioritizes speed and sensitivity to detect phosphorescence briefly brought out by UV light, using specific electronics like photodiodes and amplifiers.
  3. Experimenting with common household materials showed that some items, like powdered milk and natural ruby, exhibit weak phosphorescence, although many other items didn't show any notable glow.

Will Oxford Nanopore Sue Roche?

ASeq Newsletter โ€ข 14 implied HN points โ€ข 25 Feb 25
๐Ÿ”ฌ Science Biotechnology Intellectual Property Legal issues Genetics Research
  1. Oxford Nanopore has been actively suing companies like BGI without concrete claims of intellectual property infringement. This raises questions about their legal strategy.
  2. Roche might also face issues if they used Oxford's sequencers, especially since there are user agreements involved that they might have overlooked.
  3. Oxford Nanopore has a history of using the hairpin duplex approach, which led to legal disputes in the past. It will be interesting to see if their current patents hold up against previous art from competitors.

๐—ผ๐Ÿฏ โ€œ๐—”๐—ฅ๐—– ๐—”๐—š๐—œโ€ ๐—ฝ๐—ผ๐˜€๐˜๐—บ๐—ผ๐—ฟ๐˜๐—ฒ๐—บ ๐—บ๐—ฒ๐—ด๐—ฎ๐˜๐—ต๐—ฟ๐—ฒ๐—ฎ๐—ฑ: ๐˜„๐—ต๐˜† ๐˜๐—ต๐—ถ๐—ป๐—ด๐˜€ ๐—ด๐—ผ๐˜ ๐—ต๐—ฒ๐—ฎ๐˜๐—ฒ๐—ฑ, ๐˜„๐—ต๐—ฎ๐˜ ๐˜„๐—ฒ๐—ป๐˜ ๐˜„๐—ฟ๐—ผ๐—ป๐—ด, ๐—ฎ๐—ป๐—ฑ ๐˜„๐—ต๐—ฎ๐˜ ๐—ถ๐˜ ๐—ฎ๐—น๐—น ๐—บ๐—ฒ๐—ฎ๐—ป๐˜€

Marcus on AI โ€ข 8378 implied HN points โ€ข 22 Dec 24
๐Ÿ•น Technology AI Software Data Research Innovation
  1. Many experts feel that the recent test called ARC-AGI should not have been labeled as such. It wasn't a proper test for Artificial General Intelligence.
  2. The presentation was confusing and didn't clearly show what the AI was tested on. This left people with the impression that the AI performed better than it actually did.
  3. There's a need for more scientific scrutiny of the results. Until we get that, we can't really compare the AI's performance fairly with humans.
Get a weekly roundup of the best Substack posts, by hacker news affinity:

The Sequence Engineering #556: Inside Anthropic's New Open Source AI Interpretability Tools

TheSequence โ€ข 49 implied HN points โ€ข 04 Jun 25
๐Ÿ•น Technology AI Interpretability Open Source Research Development
  1. Anthropic is becoming a leader in AI interpretability, which helps explain how AI systems make decisions. This is important for understanding and trusting AI outputs.
  2. They have developed new tools for tracing the thought processes of language models, helping researchers see how these models work internally. This makes it easier to improve and debug AI systems.
  3. Anthropic's recent open source release of circuit tracing tools is a significant advancement in AI interpretability, providing valuable resources for researchers in the field.

Take The 2025 ACX Survey

Astral Codex Ten โ€ข 7157 implied HN points โ€ข 20 Dec 24
๐ŸšŒ Education Surveys Psychology Research Findings Incentives
  1. There's a reader survey available that helps understand who reads the blog and explore psychological findings. It's like a fun way to learn more about people's interests.
  2. Taking the survey will take around 20 to 30 minutes, and participants have a chance to win a free one-year subscription. It's a nice incentive to get more people involved.
  3. The survey closes on January 5, so it's important to fill it out before then to make your voice heard.

Achieving Government Efficiency Requires More, Not Less, Investment

The Good Science Project โ€ข 44 implied HN points โ€ข 18 Feb 25
๐Ÿ‡บ๐Ÿ‡ธ U.S. Politics Government Efficiency Public Policy Research Investment
  1. Government efficiency means getting more value from every dollar spent, not just cutting costs. Investing wisely can help achieve better results.
  2. To make government work better, we need to invest in research and development. This can help find smarter ways to deliver public services that benefit everyone.
  3. Cutting government spending isn't the answer to making it more efficient. We need to assess which programs work best and invest in them to create more social value.

Weekly Top Picks #96

The Algorithmic Bridge โ€ข 276 implied HN points โ€ข 03 Feb 25
๐Ÿ•น Technology Artificial Intelligence Machine Learning Automation Geopolitics Research
  1. OpenAI has launched two new AI agents, Operator and Deep Research, which focus on web tasks and detailed reports. Deep Research is particularly useful right now.
  2. OpenAI's o3-mini model is now free and demonstrates strong reasoning capabilities. This shows that powerful AI tools can be accessible to everyone.
  3. AI technology is evolving rapidly, and companies can benefit collectively from its advancements. Telling an AI to think longer can actually improve its performance.

Last Week in AI #291 - Nobel Prizes, Adobe's Video Model, Tesla's Robots

Last Week in AI โ€ข 99 implied HN points โ€ข 16 Oct 24
๐Ÿ•น Technology Artificial Intelligence Software Development Research Innovation Automation
  1. Two scientists won a Nobel Prize in Physics for their important work on artificial intelligence and neural networks, showing how AI is changing technology and society.
  2. Adobe has released a new AI video model that helps users create and edit videos easily, bringing exciting tools to programs like Premiere Pro.
  3. Tesla showcased new robots and vehicles at an event, but some people felt the demonstrations weren't as impressive as expected, leading to a decline in Tesla's stock.

Yes, scientific progress depends on like a thousand people

The Intrinsic Perspective โ€ข 9247 implied HN points โ€ข 29 Oct 24
๐Ÿ”ฌ Science Mentorship Innovation Research
  1. Scientific progress relies heavily on mentorship and connections within a large academic network. Many successful scientists come from the same elite lineage.
  2. Certain families of scientists are particularly influential, as seen with many Nobel Prize winners tracing back to a small number of academic mentors.
  3. While some critique the system as elitist, having a productive cultural substructure in science can be beneficial, and it's important to support this environment for future breakthroughs.

o3, AGI, the art of the demo, and what you can expect in 2025

Marcus on AI โ€ข 6481 implied HN points โ€ข 21 Dec 24
๐Ÿ•น Technology AI Software Ethics Research Economics
  1. OpenAI's new model, o3, was shown in a demo, but we can't be sure yet if it truly represents advanced AI or AGI. The demo only highlighted what OpenAI wanted to show and didn't allow public testing.
  2. The cost of using o3 is really high, potentially making it impractical compared to human workers. Even if it gets cheaper, there are concerns about how effective it would be across different tasks.
  3. Many claims about reaching AGI might pop up in 2025, but those claims need to be taken with caution. True advances in AI should involve solving more foundational problems rather than just impressive demos.

โš›๏ธ๐ŸŒŸ Our fusion future: A Quick Q&A with โ€ฆ nuclear engineer Michael Ford

Faster, Please! โ€ข 456 implied HN points โ€ข 21 Jan 25
๐Ÿ”ฌ Science Energy Engineering Physics Technology Research
  1. Fusion energy has great potential and is seen as safer than fission because it doesn't have the same risks of catastrophic accidents.
  2. Many companies are investing in fusion technology, which is leading to a quicker development pace than in the past.
  3. Public perception and understanding of fusion will be important for its acceptance and can affect the costs and deployment of fusion power systems.

Try The 2025 ACX/Metaculus Forecasting Contest

Astral Codex Ten โ€ข 3854 implied HN points โ€ข 20 Jan 25
๐Ÿ”ฌ Science Forecasting Predictions Technology Research Competition
  1. The 2025 ACX/Metaculus Forecasting Contest is now open for predictions. It's a great opportunity for anyone interested to share their forecasts on various topics.
  2. This year, there are new forecasting bots participating, and it'll be exciting to see how they compare to top human forecasters. The contest wants to explore how well these bots can predict outcomes.
  3. The questions this year are designed to be interesting and relevant, so many people can take part. The contest aims to engage everyone's thoughts on important issues.

AlphaGeometry2: Impressive accomplishment, but still a long path ahead

Marcus on AI โ€ข 3161 implied HN points โ€ข 17 Feb 25
๐Ÿ•น Technology Artificial Intelligence Machine Learning Computer Science Mathematics Research
  1. AlphaGeometry2 is a specialized AI designed specifically for solving tough geometry problems, unlike general chatbots that tackle various types of questions. This means it's really good at what it was built for, but not much else.
  2. The system's impressive 84% success rate comes with a catch: it only achieves this after converting problems into a special math format first. Without this initial help, the success rate drops significantly.
  3. While AlphaGeometry2 shows promising advancements in AI problem-solving, it still struggles with many basic geometry concepts, highlighting that there's a long way to go before it can match high school students' understanding in geometry.

โœจโฉ AI acceleration: the solution to AI risk

Faster, Please! โ€ข 548 implied HN points โ€ข 15 Jan 25
๐Ÿ•น Technology AI Security Research Innovation Policy
  1. AI development is racing forward, and the first to achieve superintelligence could have a big edge in power and resources.
  2. Speeding up technological progress may actually reduce risks of disasters because it limits the time we stay exposed to dangerous phases of development.
  3. We should focus on managing AI risks through better safety measures instead of slowing down its progress, as slowing down might lead to bigger problems.

The Smoking Gun

Nepetalactone Newsletter โ€ข 10633 implied HN points โ€ข 19 Jan 24
๐Ÿ”ฌ Science Research Health Impact
  1. The evidence strongly supports that COVID-19 was made in a lab.
  2. There is a debate within the community on various strategies to address pandemic-related issues.
  3. The focus should shift towards examining the origins of the virus and preventing future lab leaks.

โœจโ˜ข A (pricey) Manhattan Project for AI Safety?

Faster, Please! โ€ข 456 implied HN points โ€ข 17 Jan 25
๐Ÿ•น Technology AI safety Regulation Innovation Research Security
  1. AI safety may require a huge investment, like $250 billion, to ensure we can manage its risks effectively. This is much more than what was spent on the atomic bomb during World War II.
  2. Researchers believe that speeding up technological progress can actually help reduce risks from advanced AI. The idea is that the faster we move forward, the less time we have for potential dangers to develop.
  3. Many experts suggest that the U.S. government might need to take charge of AI development to ensure safety and security, creating a major project similar to the Manhattan Project. This would involve merging AI labs and improving defenses against foreign threats.

May 2025 Updates

New Things Under the Sun โ€ข 160 implied HN points โ€ข 27 May 25
๐Ÿ”ฌ Science Research Innovation R&D Economics Technology
  1. Recent studies show that government funding for research and development (R&D) yields high returns. Estimates suggest that every dollar spent on non-defense government R&D could generate benefits worth up to 210%.
  2. The composition of R&D funding has changed over the years. While support for applied research has dropped, funding for basic research has remained stable as a fraction of GDP.
  3. A significant research paper claiming that AI boosts innovation was retracted due to issues with the data and its credibility. This highlights the need for careful evaluation of research, especially when it could have major implications.

The Sequence Research #553: Self-Evaluating LLMs Are Here: Inside Meta AIโ€™s J1 Framework

TheSequence โ€ข 63 implied HN points โ€ข 30 May 25
๐Ÿ•น Technology AI Machine Learning Software Innovation Research
  1. LLMs are now used as judges, which is an exciting new trend in AI. This can help improve how we evaluate AI outputs.
  2. Meta AI's J1 framework is a significant development that makes LLMs more like active thinkers rather than just content creators. This means they can make better evaluations.
  3. Using reinforcement learning, J1 allows AI models to learn effective ways to judge tasks. This helps ensure that their evaluations are both reliable and understandable.

When will psychiatry invent the barcode?

Wood From Eden โ€ข 1344 implied HN points โ€ข 04 Dec 24
๐Ÿฅ Health Politics Mental health Psychiatry Diagnosis Research Data Analysis
  1. Psychiatry has a problem with labels. Many old labels have been removed without clear replacements, making research and understanding harder.
  2. Using numbers instead of words could help describe a person's mental health better. A barcode-like system could show traits and abilities at a glance.
  3. Psychology is subjective and changes over time. Collecting more data through tests can help improve understanding and research in mental health.

What Is The Current Evidence for mRNA Vaccine Shedding?

The Forgotten Side of Medicine โ€ข 9512 implied HN points โ€ข 08 Jan 24
๐Ÿฅ Health & Wellness Vaccination Symptoms Research Prevention Effects
  1. Shedding from mRNA vaccines is a real concern, with varying levels of sensitivity among individuals.
  2. Symptoms of shedding exposure often overlap with long COVID and vaccine side effects, particularly affecting women.
  3. Shedding can occur through proximity, skin-to-skin contact, and other secretions, with a range of symptoms such as menstrual abnormalities, bruising, dizziness, and more.

DeepSeek: Frequently Asked Questions

Artificial Ignorance โ€ข 243 implied HN points โ€ข 28 Jan 25
๐Ÿ•น Technology AI Software Innovation Data Research
  1. DeepSeek is a new AI company that has made a big impact by focusing on research instead of just selling products. It started quietly but became popular with its recent models that work well and are cheaper than competitors.
  2. Their latest products, DeepSeek V3 and R1, perform similarly to big names like ChatGPT but at much lower prices, making AI more accessible. People can even use their chatbot for free on their website.
  3. DeepSeek's success has raised questions about the future of AI development, suggesting that state-of-the-art systems can be built without spending billions. This shift in the industry has attracted significant attention and worry from major tech companies.

Neuroscience is pre-paradigmatic. Consciousness is why

The Intrinsic Perspective โ€ข 28015 implied HN points โ€ข 09 Jan 24
๐Ÿ”ฌ Science Neuroscience Consciousness Research Brain Theory
  1. Neuroscience is considered pre-paradigmatic due to the lack of a well-accepted theory of consciousness.
  2. Consciousness is viewed as the primary function of the brain, influencing all cognitive functions and behaviors.
  3. Anomalies in neuroscience, especially the mystery of consciousness, indicate a ripe field for a paradigm shift.

๐Ÿงช The US productivity problem that no one talks about (enough)

Faster, Please! โ€ข 1188 implied HN points โ€ข 04 Dec 24
๐Ÿ’ผ Business Economics Productivity Research Growth Innovation
  1. Weak productivity in science and research affects how well businesses do. It can slow down the growth of the economy.
  2. Many people donโ€™t talk enough about the link between research and overall productivity.
  3. Improving research productivity could help boost economic performance in the long run.

Looking Back at 2024

The Asianometry Newsletter โ€ข 3637 implied HN points โ€ข 31 Dec 24
๐Ÿ•น Technology Semiconductors Video content Research Travel Tech History
  1. The channel enjoyed a lot of growth in 2024, hitting impressive milestones with over 100 million lifetime views. It's amazing to think so many people watched the videos.
  2. Some favorite videos included stories about Nisei interpreters and Texas Instruments, which are rich in history and technology. These stories really resonate and are fun to tell.
  3. There are plans for 2025 to explore a mix of semiconductor topics and new themes to keep things fresh and engaging. It's all about balancing work and passion.

Genomic prediction of IQ is modern snake oil

The Infinitesimal โ€ข 1298 implied HN points โ€ข 06 Jul 24
๐Ÿ”ฌ Science Genetics Biology Psychology Mathematics Research
  1. Genetic tests claiming to predict IQ are not reliable. They often rely on complex methods that mostly just lead to guesswork.
  2. The accuracy of these genetic predictions is very low, explaining only a tiny fraction of variations in IQ scores. In fact, other factors like age and social environment play a much bigger role.
  3. Many of these predictions confuse people about how genetics really work. It's important to understand that these scores should be treated more like entertainment than serious assessments.

High-Throughput Chemistry and Engineering with Kadi Saar

Lever โ€ข 19 implied HN points โ€ข 24 Oct 24
๐Ÿ”ฌ Science Chemistry Biotechnology Engineering Research Health
  1. Kadi Saar has an impressive background in both chemistry and engineering. She excelled in academics and sports, even winning a talent show in mental arithmetic.
  2. Her research focuses on combining high-throughput structural biology with computational chemistry to help develop new drugs. She has shown that analyzing diverse ligand structures can lead to better drug design.
  3. Kadi emphasizes the importance of enjoying the people you work with when choosing projects. Collaborating with good people makes the journey more fulfilling.

Something weird is happening with LLMs and chess

DYNOMIGHT INTERNET NEWSLETTER โ€ข 1515 implied HN points โ€ข 14 Nov 24
๐Ÿ•น Technology AI LLMs Chess Gaming Research
  1. Large language models (LLMs) can somewhat play chess, but they struggle after the opening moves. They were not specifically designed for chess, yet they can manage to play using their text training.
  2. The performance of different language models varies significantly when playing chess. Some models like 'gpt-3.5-turbo-instruct' excel at it, while others perform very poorly.
  3. It seems that focusing on instruction tuning can make LLMs worse at chess, suggesting that training style impacts their ability to play games effectively.

A Capacitor Will Stop That... Right?

filterwizard โ€ข 19 implied HN points โ€ข 30 Sep 24
๐Ÿ•น Technology Engineering Electronics Innovation Research Education
  1. Capacitors are used to manage electrical noise and improve stability in circuits. They help smooth out fluctuations in voltage.
  2. Understanding electromagnetic compatibility (EMC) can prevent interference between electronic devices. This is important for maintaining performance and reliability.
  3. Decoupling is a key technique in design to isolate different circuit parts. It helps reduce noise and improves the overall functionality of the system.

How I Write Posts

Construction Physics โ€ข 21087 implied HN points โ€ข 19 Feb 24
๐ŸšŒ Education Writing Research Reading Editing Tools
  1. The author's writing process involves mainly two types of posts: explanation-driven and exploratory. They often write to understand complex topics themselves and then share their findings.
  2. Extensive research is a crucial part of the author's writing process. They gather information from a variety of sources like books, dissertations, and online resources.
  3. The author follows a structured approach starting from research, reading, and thinking, then moves on to compressing and structuring content, before finally drafting, editing, and posting their work.

You Don't Understand Serum Ferritin

Harnessing the Power of Nutrients โ€ข 2216 implied HN points โ€ข 23 May 24
๐Ÿฅ Health & Wellness Nutrition Iron Testing Research
  1. Ferritin, used to measure iron in blood, holds little iron but plays a crucial role, impacting health at extremes.
  2. Iron balance is key for health - too much ages, affects skin, causes diseases; too little harms brain, hormones, energy. It needs to be just right.
  3. Understanding serum ferritin's purpose is unclear; we don't know how it gets into blood or why. Limitations exist in relying solely on ferritin for iron status.

Declining trust in Zeus is a technology

Experimental History โ€ข 20553 implied HN points โ€ข 31 Jan 24
๐Ÿ”ฌ Science Research History Technology Innovation Experts
  1. Randomized-controlled trials are a relatively recent method in the history of scientific evidence.
  2. Challenging old beliefs, such as trust in ancient gods like Zeus, can lead to important scientific advancements.
  3. There is a need for more transparency and accessibility in expert knowledge to earn public trust in institutions and experts.

On hype, and the unbearable banality of ChatGPTโ€™s poetry

Marcus on AI โ€ข 4466 implied HN points โ€ข 19 Nov 24
๐Ÿ•น Technology AI Poetry Literature Critique Research
  1. A recent study claims that ChatGPT's poetry is similar to Shakespeare's, but it's important to be skeptical of such bold claims. Many experts believe the poetry is just a poor imitation, lacking genuine creativity.
  2. The critique of the AI poetry highlights that it often reads like the work of an unskilled poet who doesn't truly understand the style they're trying to emulate. This raises questions about the quality of AI-generated content.
  3. It's essential to approach AI-generated work with caution and to not get swayed by hype, as popular claims may not always reflect the true abilities of the technology.

The Sequence Research #543: The Leaderboard Illusion Challenges Chatbot Arena Type Benchmarks

TheSequence โ€ข 119 implied HN points โ€ข 16 May 25
๐Ÿ•น Technology AI Machine Learning Benchmarks Data science Research
  1. Leaderboards in AI help direct research by showing who is doing well, but they can also create problems. They might not show the whole picture of how models really perform.
  2. The Chatbot Arena is a way to judge AI models based on user choices, but it has issues that make it unfair. Some big labs can take advantage of the system more than smaller ones.
  3. To make AI evaluations better, there need to be rules that ensure fairness and transparency. This way, everyone gets a fair chance in the AI race.

AI Roundup 105: AI Action Summit

Artificial Ignorance โ€ข 54 implied HN points โ€ข 14 Feb 25
๐Ÿ•น Technology AI Regulation Innovation Geopolitics Research
  1. AI regulation is slowing down as countries disagree on how to move forward. Some leaders are critical of existing acts, leading to a lack of international agreement.
  2. China is pushing ahead in an AI arms race, pushing other countries to provide more resources for AI development. Leaders in the industry are predicting rapid advancements in AI, suggesting it might drastically change society soon.
  3. Big tech companies are making strategic partnerships and adjustments to survive in the competitive AI landscape. For example, Apple plans to work with Alibaba for AI in China while other firms are focusing on custom AI designs to reduce dependency on major chip manufacturers.

Freedom to write, freedom to research, to do something together: An interview with Ruth Ben-Ghiatย 

read โ€ข 10220 implied HN points โ€ข 31 Jul 23
๐ŸšŒ Education Academia Research Writing Income Feedback
  1. Scholars on Substack are reaching new audiences and earning income for their research and writing.
  2. Substack provides academics like Ruth Ben-Ghiat with financial freedom to pursue public-interfacing research.
  3. Academic writers use Substack for engaging with readers, testing new ideas, and shaping their research through feedback.

The โ€˜Science of Skincareโ€™ Isnโ€™t Science or Care

The Unpublishable โ€ข 10200 implied HN points โ€ข 23 Mar 23
๐Ÿฅ Health & Wellness Skincare Science Beauty Industry Research
  1. The 'Science of Skincare' focuses on marketing products, not actual care for the skin.
  2. Many skincare products are marketed as 'science-backed,' but they may not be as effective as claimed.
  3. Skin health is more about natural support like sleep, diet, and stress management than relying on skincare products.