The hottest Research Substack posts right now

And their main takeaways
Category
Top Health Politics Topics

Can donors save science?

ChinaTalk • 207 implied HN points • 12 Jun 25
🚌 Education Research
  1. Renaissance Philanthropy aims to change how funding works by acting like an investment fund, helping donors target specific goals without the heavy lifting of managing projects themselves.
  2. The organization identifies key areas in science and tech where philanthropic funding can make a real impact, especially at a time when government support is shrinking.
  3. A new model of philanthropy can encourage more donors to get involved by picking clear objectives and making the process less daunting, ultimately leading to more effective giving.

✨☢ A (pricey) Manhattan Project for AI Safety?

Faster, Please! • 456 implied HN points • 17 Jan 25
🕹 Technology Research
  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.

Are AlphaFold's new results a miracle or a mirage?

Oleg’s Substack • 37 HN points • 24 Jun 24
🔬 Science Research
  1. AlphaFold 3 can predict how drug-like molecules bind to proteins better than existing programs without needing a 3D structure of the target.
  2. Data redundancy in scientific datasets can impact the performance and interpretation of machine learning models.
  3. AlphaFold 3's occasional missed obvious insights, like atoms overlapping, raises questions about its learning methods and performance.

The Sequence Opinion #778: After Scaling: The Era of Research and New Recipes for Frontier AI

TheSequence • 28 implied HN points • 25 Dec 25
🕹 Technology Research
  1. Scaling up transformers with more data and compute drove past AI gains, but that straightforward path is hitting limits because high-quality pretraining data and scaling efficiency are finite.
  2. The field is shifting to an "age of research" where diverse experiments and new ideas, not just bigger models, will determine future breakthroughs.
  3. Progress will come from a toolbox of new recipes — like souped-up pretraining, novel architectures, and improved fine-tuning — that turn compute into faster learning, better adaptation, and fewer odd model failures.

Against Mathiness

The Honest Broker Newsletter • 1364 implied HN points • 02 Jan 24
🔬 Science Research
  1. Mathiness can allow for academic politics to masquerade as science.
  2. Good empirical research can eventually win out in truth battles, even if it takes time.
  3. It's possible to distinguish between good policy research and mathiness, despite potential professional benefits and political challenges.
Get a weekly roundup of the best Substack posts, by hacker news affinity:

OpenAI’s Deep Research: The "Big Bang" Event for AI Agents

Enterprise AI Trends • 400 implied HN points • 06 Feb 25
🕹 Technology Research
  1. OpenAI's Deep Research feature allows users to get thorough research done quickly, acting like a smart research assistant. This can save a lot of time compared to traditional searching methods.
  2. Deep Research is designed to work on its own, leading the research process instead of needing constant input. This makes it more productive and user-friendly.
  3. As Deep Research becomes popular, competition in the AI space will change. Companies will now need to clearly explain how their offerings are better than Deep Research, raising the standard for AI tools.

Was 8th Century China Meritocratic?

The Great Gender Divergence • 157 implied HN points • 10 Feb 24
💾 History Research
  1. In 8th Century China, there was a widespread belief in meritocracy and social mobility, which were endorsed by Confucius and Empress Wu Zetian.
  2. Elites might have had advantages like better education for their sons and using nepotism for promotions, which could challenge the true meritocratic nature of the system.
  3. Recent research on social mobility during the Tang Dynasty provides insights into the career paths, imperial exam pass rates, and ancestral backgrounds of individuals through tomb epitaphs.

Beyond Five Seconds: Rethinking First Impressions in UX

UX Psychology • 238 implied HN points • 08 Dec 23
🕹 Technology Research
  1. First impressions are crucial in UX and can influence user engagement throughout their interaction with an interface.
  2. Traditional 'five-second tests' in UX may not account for variations in users' cognitive abilities and the complexity of visual designs.
  3. To enhance UX design, consider customizing testing based on cognitive abilities, evaluating visual complexity, and rethinking the standard 'five-second rule.'

Let's Talk About Noom

Weight and Healthcare • 619 implied HN points • 07 Dec 22
🏥 Health & Wellness Research
  1. Noom's claims of being a brand new way to lose weight while also helping to keep weight off for good are inconsistent. Long-term efficacy cannot be proven with short-term results.
  2. Short-term weight loss statistics from Noom might not reflect sustained success, as most participants tend to regain weight within a few years, similar to other diet programs.
  3. Studies on Noom's effectiveness often involve low retention rates and short follow-up durations, which are not sufficient to support their marketing claims of long-term weight loss maintenance.

Walk in the Pines #144

Pinecone Weekly Brief • 196 implied HN points • 07 Jan 24
💰 Finance Research
  1. The post includes a list of recommended reads, listens, and watches for the week.
  2. The content covers a wide range of topics such as economics, finance, and learning.
  3. There are links provided for further exploration of the recommended content.

A short history of SHAP

Mindful Modeler • 359 implied HN points • 30 May 23
🕹 Technology Research
  1. Shapley values originated in game theory in 1953 and contributed to fair resource distribution methods.
  2. In 2010, Shapley values were introduced to explain machine learning predictions, but didn't gain traction until the SHAP method in 2017.
  3. SHAP gained popularity for its new estimator for Shapley values, unification of existing methods, and efficient computation, leading to widespread adoption in machine learning interpretation.

Prospective Memory and UX

UX Psychology • 317 implied HN points • 04 May 23
🔬 Science Research
  1. Prospective memory is the ability to remember to do something in the future. It's crucial for daily tasks like attending meetings or taking medication.
  2. There are two main types of prospective memory: event-based (triggered by cues) and time-based (triggered by specific times). Distractions or cognitive load can lead to failures in remembering to do a task at the right time.
  3. Designing interfaces with visual cues, gamification, habit formation, contextual design, and cognitive offloading can improve users' prospective memory performance and create more engaging experiences.

The Fusion Energy Landscape

Ulysses • 359 implied HN points • 05 Apr 23
🔬 Science Research
  1. Fusion energy promises clean, pollution-free energy that can be built anywhere and lasts millions of years without geopolitical disputes over fuel reserves.
  2. Different fusion technologies like magnetic confinement, inertial confinement, and magneto-inertial confinement offer diverse approaches to achieving fusion reactions with varying benefits and challenges.
  3. Developing fusion energy requires extensive funding, traditional engineering solutions, and expertise underscoring a journey to harness limitless energy potential.

Edge 360: Meet Ghostbuster: An AI Technique for Detecting LLM-Generated Content

TheSequence • 1310 implied HN points • 11 Jan 24
🕹 Technology Research
  1. Berkeley University developed a method to detect AI-generated tokens in documents using probability distribution.
  2. Ghostbuster is an AI technique for identifying AI-generated text by calculating token likelihood and using a conclusive classifier.
  3. The technique by Berkeley AI Research aims to tackle challenges in differentiating between human and AI-generated content.

This week in the science of learning, May 19, 2024

The Bell Ringer • 59 implied HN points • 19 May 24
🚌 Education Research
  1. Schools need better data on how students read and what materials they are using. This helps to improve teaching methods.
  2. There's a strong interest in sharing stories and research about learning methods. People want to hear what works and what doesn't.
  3. The community is encouraged to provide feedback on topics of interest, which can lead to more personalized and engaging discussions.

Long Covid Research Roundup

Your Local Epidemiologist • 1284 implied HN points • 18 Jan 24
🏥 Health & Wellness Research
  1. The prevalence of long Covid may be decreasing due to a decline in severe disease and the changing virus over time.
  2. Vaccines, especially with additional doses, help protect against long Covid in both children and adults.
  3. Long Covid is a complex condition with various underlying causes, such as viral persistence, immune responses, and gut microbiome links.

Mini Exasperations #2

Exasperated Infrastructures • 14 implied HN points • 17 Jan 26
🇺🇸 U.S. Politics Research
  1. The TRB annual meeting is the most valuable week for researchers and academics — go for hallway conversations and selective evening receptions, and don’t try to see everything.
  2. U.S. transit systems need massive, sustained investment (roughly $230 billion a year in one estimate) to repair decades of underfunding and remain competitive.
  3. Transit planning is full of tradeoffs like coverage versus speed, so practical changes (for example, fewer bus stops) and strong community input can free resources and improve service.

News! & Longer Living Through Chemistry

Unsettled Science • 545 implied HN points • 26 Oct 24
🔬 Science Research
  1. A new Substack newsletter is being launched due to challenges faced in a joint project. Leaving the previous newsletter was not an easy decision but necessary to move forward.
  2. There is a mention of an anti-aging pill in the news. It's important to think carefully about whether to take such a pill or not.
  3. The focus is on exploring longer life through chemistry. This suggests that science might hold keys to better health and longevity.

The Sequence Radar #688: The Transparent Transformer: Monitoring AI Reasoning Before It Goes Rogue

TheSequence • 154 implied HN points • 20 Jul 25
🕹 Technology Research
  1. AI researchers are exploring a way to monitor advanced AI reasoning to catch any dangerous behavior early. This method looks at how AI models 'think' through problems using something called chains of thought.
  2. This monitoring method is helpful but can be fragile. As AI models get better, they might stop using natural language reasoning, making it harder to understand their thought processes.
  3. There is a big push for more research to keep this monitoring effective. By establishing clear benchmarks, we can better evaluate and improve how we observe AI reasoning.

Happy Llama Day 🦙🤩 ✨

Sector 6 | The Newsletter of AIM • 79 implied HN points • 20 Apr 24
🕹 Technology Research
  1. Meta launched Llama 3, an advanced open-source language model that outshines its competitors in reasoning and coding tasks. This model is creating a lot of buzz for its performance.
  2. Andrej Karpathy, a former OpenAI scientist, is very excited about Llama 3 and thinks it will be a strong competitor against GPT-4.
  3. Llama 3 is designed with a massive 400 billion parameters, making it a powerful tool for various applications in AI.

Letters from the field #5 - Season 2 announcement + Call for participation

On Looking • 79 implied HN points • 19 Apr 24
🎨 Art & Illustration Research
  1. Season 2 of the research will focus on exploring the perspectives of designers, art directors, editors, and production managers in the creative industry.
  2. The research involves conducting in-person interviews as well as observations within magazines, design studios, and advertising agencies in New York to understand how creative professionals work and make decisions.
  3. The researcher is seeking New York-based creatives to participate in the study, emphasizing the importance of observations in workspaces to gather insights for the research.

Deep Into the Depths

Atlas of Wonders and Monsters • 339 implied HN points • 27 Feb 25
🕹 Technology Research
  1. AI tools have started using the term 'deep' to suggest they dig into more complex information, but this may often not be the case. Many still just skim the surface instead of really exploring.
  2. While AI is getting better at research by gathering information quickly, true deep research requires more human-like exploration and understanding. It's about going beyond just looking up facts.
  3. Don't be fooled by the hype around AI's 'deep research' capabilities. They are useful, but they aren't as profound or groundbreaking as some might claim.

Weight Loss Surgery and Kids Part 2

Weight and Healthcare • 299 implied HN points • 12 Aug 23
🏥 Health & Wellness Research
  1. The study on weight loss surgeries did not focus on pediatric populations, making its relevance to children in question. The short follow-up periods in the study are concerning for assessing long-term implications of surgeries.
  2. The study authors' potential conflicts of interest raise concerns about the objectivity of their research, especially when making recommendations for surgical treatments, including for children.
  3. The study's methodology and statistical analysis may not present a comprehensive view of the risks and effectiveness of weight loss surgeries, potentially downplaying negative outcomes and psychological impacts.

The World-as-History

The Oswald Spengler Project • 59 implied HN points • 15 May 24
💾 History Research
  1. The post discusses the translation of Spengler's works into English, particularly focusing on his intention for the academic journal 'The World as History' and the journal's editor Hans Erich Stier.
  2. The journal 'The World as History' aimed to publish historical research aligned with Spengler's distinct methodology.
  3. The post hints at a possible revolution in cartography, linking to further content on the subject.

Import AI 328: Cheaper StableDiffusion; sim2soccer; AI refinement

Import AI • 339 implied HN points • 08 May 23
🕹 Technology Research
  1. Training image models can be cheaper with smart tweaks like Low Precision GroupNorm and Low Precision LayerNorm. Companies like Mosaic are leading the way in AI industrialization.
  2. Prominent AI researcher Geoff Hinton has expressed concerns about the rapid progress and control of advanced AI models. His departure from Google highlights the growing worries in the field.
  3. New companies like Lamini are offering services to fine-tune existing AI models, indicating further industrialization of AI. Startups like these are bridging the gap between AI products and consumers.

Education is a Joke

Pershmail • 294 implied HN points • 13 Jul 23
🚌 Education Research
  1. Education can be seen as a joke, focusing on entertaining instead of teaching.
  2. Educational conferences and events are often criticized for prioritizing entertainment over substance.
  3. There is a growing concern about the unseriousness and anti-intellectualism in the field of education.

"Think of the Implications of Publishing"

The Honest Broker Newsletter • 1707 implied HN points • 17 Jul 23
🔬 Science Research
  1. A whistleblower exposed corruption in peer review in climate science, leading to the retraction of a paper.
  2. The process revealed how activist scientists influenced the publishing process behind the scenes.
  3. The case highlights ongoing challenges in climate science and the politicization of scientific publishing.

The CAP Theorem of Clustering: Why Every Algorithm Must Sacrifice Something

Confessions of a Code Addict • 529 implied HN points • 29 Oct 24
🕹 Technology Research
  1. Clustering algorithms can never be perfect and always require trade-offs. You can't have everything, so you have to choose what matters most for your project.
  2. There are three key properties that clustering should ideally have: scale-invariance, richness, and consistency, but no algorithm can achieve all three simultaneously.
  3. Understanding these sacrifices helps in making better decisions when using clustering methods. Knowing what to prioritize can lead to more effective data analysis.

Welcome to Yokohama and ICRA 2024!

Robots & Startups • 59 implied HN points • 12 May 24
🕹 Technology Research
  1. The average technical recruiter is young, and studies suggest they may favor younger job candidates over older ones.
  2. ICRA 2024 in Yokohama is a large event with over 5000 roboticists and 120+ robotics companies showcasing quadruped and humanoid robots.
  3. Academic robotics conferences like ICRA feature a variety of activities beyond paper presentations, including robotics competitions, industry exhibitions, career fairs, debates, and more.

New Record(ings)!

Brain Blast • 39 implied HN points • 12 Jun 24
🔬 Science Research
  1. Intra-cranial electroencephalography (iEEG) is a technology using tiny electrodes in the brain to record signals, and Precision Neuroscience set a world record with 4,096 electrodes.
  2. Wilder Penfield pioneered important techniques in brain surgery and epilepsy treatment, including the Montreal Procedure and temporal lobectomy, affecting both surgical outcomes and scientific understanding of the brain.
  3. Stereo EEG, a variant of iEEG, allows for deep brain signal recording, and using iEEG in a Stroop task helped identify brain regions involved in attentional control processes in the prefrontal cortex.

On Plagiarism and Related Issues

Imperfect Information • 157 implied HN points • 24 Jan 24
🚌 Education Research
  1. Plagiarism detection tools are widespread and incentives are strong to uncover copied content.
  2. Different types of plagiarism exist, from accidental use of others' work to theft of novel ideas.
  3. Plagiarism war may lead to accusations of minor transgressions, but may not detect serious intellectual misconduct.

Walk in the Pines #146

Pinecone Weekly Brief • 157 implied HN points • 21 Jan 24
💰 Finance Research
  1. The post shares information about 'Walk in the Pines #146.'
  2. It features Charts, Reads, Listens, and Watches of the Week.
  3. The post discloses that Pinecone Macro Research, LLC is an independent research firm.

Two Of My Favorite Investments: Chris DeMuth

QTR’s Fringe Finance • 14 implied HN points • 15 Jan 26
💰 Finance Research
  1. Ask “what would this look like if it were easy?” to reframe problems and find elegant, high-leverage solutions instead of forcing effort and complexity.
  2. Look for smaller competitors with close comparables that show a clear path to closing the market-cap gap with larger incumbents; those situations can produce big, relatively straightforward gains.
  3. Using comps to visualize an easy outcome helps identify investments where scaling feels likely and the upside is easy to justify.

Fixing Higher Ed: The Bailey-Caplan Convo

Bet On It • 105 implied HN points • 22 Aug 25
🚌 Education Research
  1. There's concern about the increasing politicization of universities, which some liken to past political eras like McCarthyism. This raises questions about how knowledge and education are being influenced by ideology.
  2. Many feel that government funding for higher education may actually be contributing to wasted time for students. It's important to think about how much funding should be given and how it should be used.
  3. The debate about what to save in higher education includes discussions on faculty hiring and the importance of ideological diversity. Ensuring a range of viewpoints can help foster a more open inquiry in academic settings.

Talking science communication with a Mythbuster

Science Forever • 119 implied HN points • 26 Feb 24
🔬 Science Research
  1. Science communication should be personalized and emotional to engage people effectively.
  2. Revisiting and retesting myths is an essential part of the scientific process and can lead to different results.
  3. Encouraging transparency in scientific publishing, advocating for open access, and embracing the self-correcting nature of science can help address the reproducibility crisis.

Space Ambition 1st Anniversary: We Got a Bit Closer to the Stars

Space Ambition • 279 implied HN points • 15 Sep 23
🔬 Science Research
  1. Space Ambition attracted over 8,000 readers and completed more than 50 research projects in its first year. It's great to see such a big community focused on space and technology!
  2. The team believes that space technology can solve many of Earth's problems and aims to make it more understandable and accessible for everyone. They want to connect space with everyday life.
  3. Readers have shared inspiring stories about how the blog has influenced their careers, which shows the positive impact of their work in the space sector. It's awesome to see people motivated to pursue their dreams!