Longevity InTime: Autonomous AI Institute. Anti-Aging Digital Health Immortality Transhumanist AI Channel
Kanalga Telegram’da o‘tish
NVIDIA inception Member, Nebius AI Discovery Awards semifinalist Potentially first $1T Longevity BioTech AI company Part of Longevity Ecosystem LongevityInTime.com @PickleballPartners Shop https://web.tribute.tg/l/lr Homes www.Africa.Villas
Ko'proq ko'rsatish1 040
Obunachilar
Ma'lumot yo'q24 soatlar
-117 kun
-1930 kun
Ma'lumotlar yuklanmoqda
O'xshash kanallar
Taglar buluti
Kirish va chiqish esdaliklari
---
---
---
---
---
---
Obunachilarni jalb qilish
Noyabr '26Noy '26
Noyabr '260
0 kanalda
Oktabr '26
+21
0 kanalda
Get PRO
Sentabr '26
+9
0 kanalda
Get PRO
Avgust '26
+20
0 kanalda
Get PRO
Iyul '26
+13
0 kanalda
Get PRO
Iyun '26
+14
0 kanalda
Get PRO
May '26
+34
0 kanalda
Get PRO
Aprel '26
+36
0 kanalda
Get PRO
Mart '26
+14
0 kanalda
Get PRO
Fevral '26
+14
0 kanalda
Get PRO
Yanvar '26
+15
0 kanalda
Get PRO
Dekabr '25
+12
0 kanalda
Get PRO
Noyabr '25
+14
0 kanalda
Get PRO
Oktabr '25
+18
0 kanalda
Get PRO
Sentabr '25
+21
0 kanalda
Get PRO
Avgust '25
+32
0 kanalda
Get PRO
Iyul '25
+114
0 kanalda
Get PRO
Iyun '25
+9
0 kanalda
Get PRO
May '25
+6
0 kanalda
Get PRO
Aprel '25
+5
0 kanalda
Get PRO
Mart '25
+9
0 kanalda
Get PRO
Fevral '25
+9
0 kanalda
Get PRO
Yanvar '25
+20
0 kanalda
Get PRO
Dekabr '24
+14
0 kanalda
Get PRO
Noyabr '24
+8
0 kanalda
Get PRO
Oktabr '24
+7
0 kanalda
Get PRO
Sentabr '24
+9
0 kanalda
Get PRO
Avgust '24
+10
0 kanalda
Get PRO
Iyul '24
+15
0 kanalda
Get PRO
Iyun '24
+16
0 kanalda
Get PRO
May '24
+13
0 kanalda
Get PRO
Aprel '24
+10
0 kanalda
Get PRO
Mart '24
+13
0 kanalda
Get PRO
Fevral '24
+17
0 kanalda
Get PRO
Yanvar '24
+14
0 kanalda
Get PRO
Dekabr '23
+14
0 kanalda
Get PRO
Noyabr '23
+10
0 kanalda
Get PRO
Oktabr '23
+8
0 kanalda
Get PRO
Sentabr '23
+14
0 kanalda
Get PRO
Avgust '23
+24
0 kanalda
Get PRO
Iyul '23
+7
0 kanalda
Get PRO
Iyun '23
+9
0 kanalda
Get PRO
May '23
+10
0 kanalda
Get PRO
Aprel '23
+14
0 kanalda
Get PRO
Mart '23
+22
0 kanalda
Get PRO
Fevral '23
+20
0 kanalda
Get PRO
Yanvar '23
+7
0 kanalda
Get PRO
Dekabr '22
+6
0 kanalda
Get PRO
Noyabr '22
+17
0 kanalda
Get PRO
Oktabr '22
+17
0 kanalda
Get PRO
Sentabr '22
+9
0 kanalda
Get PRO
Avgust '22
+11
0 kanalda
Get PRO
Iyul '22
+2
0 kanalda
Get PRO
Iyun '22
+1
0 kanalda
Get PRO
May '22
+3
0 kanalda
Get PRO
Aprel '22
+8
0 kanalda
Get PRO
Mart '22
+4
0 kanalda
Get PRO
Fevral '220
0 kanalda
Get PRO
Yanvar '220
0 kanalda
Get PRO
Dekabr '210
0 kanalda
Get PRO
Noyabr '210
0 kanalda
Get PRO
Oktabr '210
0 kanalda
Get PRO
Sentabr '210
0 kanalda
Get PRO
Avgust '210
0 kanalda
Get PRO
Iyul '210
0 kanalda
Get PRO
Iyun '21
+73
0 kanalda
Get PRO
May '21
+13 445
0 kanalda
Get PRO
Aprel '21
+10 157
0 kanalda
Get PRO
Mart '21
+18 260
0 kanalda
Get PRO
Fevral '21
+41 397
0 kanalda
| Sana | Obunachilarni jalb qilish | Esdaliklar | Kanallar | |
| 02 Noyabr | 0 | |||
| 01 Noyabr | 0 |
Kanal postlari
Impact of Anti-Regulatory Biotech Message Reaches Mainstream
At least 200 people have engaged with the critique of draconian regulations and agencies like the FDA. This pushes the anti-regulatory stance in biotech closer to the mainstream.
Every viral tweet moves Transhumanism/Immortalism/Longevity from a niche topic to a universal one. That is the goal: more memes, more advocacy, and saying "Transhumanism" at every turn.
🔗 Source: @solid_state_humanity
| 2 | Personal milestone: first viral Twitter post hits 220 likes, 16.8K views
The post reached 220 likes and 16.8K views on X — a first for me. The image is somewhat relevant to our topic, though not deeply conceptual.
🔗 Read original → | 7 |
| 3 | AI Will Let Small Teams Evaluate More Drug Ideas, Investor Simon Barnett Predicts
On October 30, Dimension partner Simon Barnett proposed outsourcing information gathering and external lab coordination to AI. In an interview with Endpoints News journalists Drew Armstrong and Andrew Dunn, he discussed how this shifts scientific workflows and who funds the experiments.
The catalyst was AI developer Anthropic's acquisition of biotech Mithril Medicine, a company with a single full-time employee. At Mithril, the AI assistant Claude drafted correspondence to labs while the human edited and sent it.
Pharma has long spun out development to project-specific companies; in 2011, Eli Lilly licensed a migraine drug candidate to Arteaus, buying it back after clinical trials in 2014. In this model, the small firm pays for experiments conducted by external lab staff.
Barnett suggests pairing an experienced drug developer with an AI specialist. The AI would track competitors, run specialized protein-structure tools, and coordinate outside labs — citing voice AI agents that call facilities directly. He argues salary savings "do not determine the success or failure of this model."
Instead, freed from coordination burdens, teams could parallelly evaluate more related ideas at early stages — for instance, multiple ways to activate cellular receptors. The developer assesses experimental suitability for each idea; the AI specialist selects tools and identifies data gaps. Together, they decide which ideas to kill and which to fund.
Barnett sees a commercial niche for such firms: projects with markets too small for big pharma. Investor Bruce Booth, in an essay discussed during the interview, predicts AI will ease molecule creation and swell the pipeline for popular targets. Developers of similar drugs will compete for lab slots, trial sites, and patients — slowing patient enrollment.
Barnett warns that undifferentiated projects will struggle to raise capital. Experiments need lab animals; human trials need volunteers. Asked if supply can meet rising demand, he replied: "We have a limited number of monkeys."
🔗 Read original → | 6 |
| 4 | Brain Preservation After Death as Alternative to AI Race with Extinction Risk
Nate Soares, president of the Machine Intelligence Research Institute (MIRI), debated Vladimir Shakirov on November 1 about the cost of pausing AI development. Shakirov highlighted people dying from aging and disease, while Soares argued for weighing all future generations. Founded in 2000 to accelerate superhuman AI, MIRI now urges a halt until the alignment problem is solved.
The discussion opened with conditions for resuming development. Soares proposed a benchmark: resume only when extinction risk from AI is comparable to the risk of losing humanity's future from other causes. Shakirov countered that aging and disease mortality create urgency; a prolonged pause risks patients dying before cures arrive, giving them reason to accelerate AI for medical breakthroughs.
Soares objected that extinction deprives both the living and all possible descendants. He views aging mortality as insufficient grounds for such a gamble. For those wanting to reach future medicine, he endorses brain preservation after legal death. Long-term memory and personality traits depend on neural connectivity; preservation aims to maintain these connections until restoration technologies exist.
In August, neuroscientist Ariel Zeleznikow-Johnston linked brain preservation to reducing AI race dynamics. His argument: if people have another path to future medicine, they have less incentive to rush AI for personal survival. In the current debate, Soares advocated preservation for those who choose it: "Reach the future without a reckless bet on everyone else's lives."
🔗 Read original → | 5 |
| 5 | One Manager Could Oversee Multiple AI-Driven Drug Projects, Entrepreneur Proposes
In an October 30 essay, Igan Peltan distinguishes two causes of slow drug development: lengthy validation and organizational delays. Discussant Jesse Brodkin argues for investing first in animal experiments that better predict human outcomes. Peltan suggests philanthropic funding for this work.
Peltan is building a startup to maintain physical function in the elderly. His core question: how to know if a new molecule will help humans? Developers use animal models to study disease-linked processes and test treatments, but Peltan stresses these often poorly predict human effects. Testing a dementia preventive requires long observation to see if disease develops. Safety verification also takes time: dosing twelve men for one month cannot yield year-long safety data.
Peltan expands on drug developer Cody Tranbarger's idea of running multiple projects with AI. Additional delays stem from funding and coordination among specialists holding fragmented data. Peltan sees an AI role: aggregating information and coordinating work, freeing the lead for scientific decisions. Contract labs run experiments; multiple projects can share overhead. If AI handles more coordination, one lead can manage several projects in parallel — switching to others while awaiting results.
In Peltan's scheme, specialized AI assistants read and update a shared journal of data and decisions, while a separate assistant aligns their work with the lead. Peltan describes his own experience: when summarizing long scientific and manufacturing documents in chunks, AI can produce conclusions without source citations, forcing re-verification. He asks how assistants can exchange information while preserving the source of every claim. Humans must be able to trace any conclusion back to the original document.
Brodkin urges investment in improving the animal models used to select future drugs: "Build labs in the desert, not data centers." In reply, Peltan agrees with the model assessment, does not expect AI to impact this problem soon, and asks whether philanthropists could fund more accurate models.
Peltan speculates that AI companies pursue drug development partly for the experience of human–AI collaboration. Here, validation is slow, data is scattered across specialists, and decisions are interdependent. Similar coordination and data-provenance challenges exist in auto manufacturing or building desalination plants; lessons from a small drug company could transfer to those industries.
🔗 Read original → | 6 |
| 6 | Mathematicians Decode AI-Found Hadamard Matrices and Debate AI's Role in Research
A November 1 New Yorker profile by Rivka Galchen spotlights mathematician Terence Tao's concerns about training researchers when AI solves in hours problems that once took graduate students months. In July, Tao analyzed an AI-discovered formula that disproved a mathematical conjecture. His explanation starts with a product of three linear factors: one is isolated while the other two are multiplied. Selecting a different factor yields a new polynomial pair with the same product — a technique colleagues can now apply elsewhere.
Mathematician Geordie Williamson welcomed Tao's breakdown but reacted differently to the August publication of twelve Hadamard matrices. These are square tables of +1 and −1 where any two distinct rows match in exactly half their positions. The published sign strings and code allowed verification but not the discovery method. Williamson feared colleagues would get answers without the ideas found along the way.
In October, another group established how all twelve tables are assembled from smaller blocks. Programs for analysis and verification were written by the language model Claude. Analyzing three matrices yielded a theorem with general assembly conditions for such matrices. The authors manually verified and rewrote the theorem; its conditions enable selecting blocks for new matrices.
Mathematicians John Urschel and his MIT colleague Shah directed model work themselves. They studied Gaussian elimination — a method for solving linear systems where intermediate numbers can grow drastically, amplifying rounding errors. To assess maximum possible growth, they built special examples. In one construction, they fed models intermediate ideas from a research plan developed since 2022. The final proofs were written by the authors, who meticulously separated human and machine contributions per section. Urschel found working with models brought "no enjoyment"; he prefers finding solution paths himself.
Mathematician Charles Fefferman suggests imagining an idea with a 5% success chance that AI can verify in two days instead of a lifetime. Cheap verification, he predicts, will make humans "bolder and more inventive."
🔗 Read original → | 6 |
| 7 | Evaluate All Math Papers as If AI Helped, Says Frank Calegari
University of Chicago professor Frank Calegari proposed restructuring the mathematics profession on November 1. His plan links career recognition to problem selection, proof comprehension, and developing new research directions.
The trigger was OpenAI publishing hundreds of mathematical manuscripts on October 6. Calegari describes a growing gap: new proofs can now be generated faster than humans can grasp their ideas and connections to broader mathematics.
Previously, a difficult proof signaled mastery. If AI can produce it, a single paper no longer reveals how much the author truly understands. Calegari suggests evaluating this understanding separately. "Evaluate any work as if it were done with AI." Authors must explain why the problem matters, how it connects to other math, and what new questions it opens.
He would credit a researcher who selects a worthy problem, gets the full proof from AI, deeply understands it, and spins off new directions. Novel proof ideas also deserve recognition, even if machines could reach the same conclusion.
Since mathematicians compete for jobs, grants, and publications, Calegari wants hiring and editorial decisions to reward understanding and new questions. He advocates in-person interviews for PhD admissions and postdoc hiring. Candidates must discuss diverse areas, explain their methods, and articulate where they want to take research. Calegari prioritizes depth of knowledge over verbal fluency; the conversation tests comprehension.
He asks journals to focus more on ideas and clarity of exposition. If editorial boards keep old selection criteria, he suggests a temporary boycott of journal publication so every paper is judged on content. Calegari himself sits on the board of Essential Number Theory, which encourages clear exposition of others' results to help fellow researchers.
These standards reshape graduate training. Calegari describes the traditional path: master a technique on a variation of an old problem, then apply it broadly. Now he proposes combining independent problem-solving with a broad mathematical survey. After a course, students should grasp what a topic enables in other fields and remember those connections, even if technical details fade.
Students should learn to discuss math with peers in seminars and conversations. Calegari suggests structuring seminars around a big question: where researchers are stuck, what paths they have tried, and where to go next. Speakers may present their own new results at the end.
🔗 Read original → | 6 |
| 8 | Reducing biological age must link to health gains, says aging researcher Jesse Poganik
In an October 30 interview with biologist Eleanor Sheekey, Harvard Medical School's Poganik outlines two roles for aging clocks: predicting disease and evaluating treatments. He stresses tracking how long a clock shift persists after an intervention ends.
Epigenetic clocks gauge age via chemical marks on DNA. Poganik recalls first testing how quickly clocks respond to drugs. After one week of treatment in mice, readings were unchanged, so the team turned to stronger physiological perturbations. Researchers joined the circulatory systems of young and old mice, then separated them for recovery; this parabiosis experiment was published in a 2023 study. As a control, two young mice were joined to isolate the partner's age effect from the surgery itself.
The clocks' age estimate for young mouse tissues was compared to their chronological age. After contact with an old partner, the gap widened; after separation and recovery, it returned to baseline. Sheekey asks how long a reduced age reading lasts post-intervention. One of Poganik's criteria for true age reversal is "persistence of effect after cessation of intervention." In his proposed definition, the score must drop below baseline and stay there after the stimulus stops. Pre- and post-measurements use identical features, allowing quantification of both magnitude and duration.
A biomarker is a measurable indicator of physiological state. To test if it predicts disease risk, researchers correlate baseline readings with future illness and functional independence. Such prediction can guide screening or trial enrollment. In an Aging Biomarker Consortium paper co-authored by Poganik, the authors argue: "Useful biomarkers can be built without detailed understanding of causal links." To judge treatment benefit from a clock shift, one must separately verify that the drug-induced change predicts actual health improvement.
Poganik highlights the value of measuring a defined panel of blood proteins with a single assay. Such data integrate more easily across studies, and tests built on them validate better in new cohorts. Using proteins specific to heart, liver, or other organs yields separate organ-age scores. From one plasma sample, researchers obtain multiple scores and correlate them with organ function and disease.
In his own lab, Poganik studies already-approved drugs. Disease itself can raise age scores, while treatment lowers them. He aims to disentangle how much of the reduction reflects recovery from illness versus a direct drug effect on aging processes.
🔗 Read original → | 5 |
| 9 | L.E.K. Advises Pharma to Start Aging Drug Work with Small Teams and Partnerships
On October 29, 2026, consulting firm L.E.K. Consulting recommended that pharmaceutical companies build research experience before acquiring drug rights or entire developers. Firms can begin with joint research or a small equity stake, scaling investment as data accumulates.
Because a single aging mechanism may drive multiple diseases, a drug targeting it could have several applications. L.E.K.’s strategy starts with a specific disease — the first indication. For this indication, teams must define trial participants, duration, and the patient benefit metric regulators will evaluate. “The first indication must satisfy pharma’s commercial requirements independent of additional opportunities related to aging impact.”
Before trials, teams need to select a molecular target — such as a protein — and validate its role in the disease through experiments. While many researchers access human health databases, L.E.K. says the advantage goes to companies whose scientists can turn that data into testable hypotheses, design experiments, and collaborate with technology developers. “Building these skills and connections takes years.”
To gain this experience, L.E.K. suggests creating a small research group within an existing therapeutic area, such as cardiovascular disease. Novartis’s DARe (Diseases of Aging and Regenerative Medicine) group exemplifies this: its tissue regeneration approach originated in joint and tendon research and is now being applied to Alzheimer’s disease.
Collaboration can also extend to measuring treatment benefit. In the THRIVE project, a Stanford University-led coalition pools data from multiple organizations to build a composite metric of physical and cognitive function. Using longitudinal medical records, the team aims to identify early changes that predict chronic disease; a one-year lifestyle intervention program will test the metric’s accuracy and responsiveness. L.E.K. urges pharma to join such efforts to learn how to validate and apply novel endpoints during development.
Consultants expect results from several drug trials in 2027–2028 and anticipate successful programs will become more expensive. They advise defining in advance which results would justify increased investment. Early partnerships can give teams experience selecting and testing their own hypotheses before committing to an acquisition.
🔗 Read original → | 4 |
| 10 | AI-Accelerated Neuron Replacement: Patrick Mineo's Roadmap
On October 30, 2026, neuroscientist Patrick Mineo traced the evolution from 1980s fetal tissue grafts to modern stem-cell therapies for neuron replacement.
On March 6, Japan granted conditional, time-limited approval for a dopamine-neuron product to treat Parkinson's motor symptoms when drugs fail. In Parkinson's, dopamine-producing neurons die; transplanted immature cells are meant to mature and restore dopamine supply. For drug-resistant epilepsy originating in the temporal lobe, Neurona Therapeutics is testing inhibitory neurons that integrate into local circuits to suppress hyperexcitability.
Early grafts used variable human fetal tissue. Stem cells enable scalable, defined neuron production. A 2011 study generated dopamine neurons by mimicking midbrain developmental signals; these survived and reversed symptoms in Parkinsonian animals. Selecting the right maturity stage proved critical: too immature risks overgrowth, too mature reduces engraftment. Batches are now screened for residual stem cells that could form tumors.
Aspen Neuroscience presented a 2025 model predicting rat engraftment from gene-activity profiles. Other algorithms use imaging to guide robotic culture processing. Their manufacturing run lasted over 22 months; operators spent over 6,000 hours evaluating it. Delivery precision also matters: in the TransEuro trial, one patient received cells outside the target zone bilaterally, so predictive models must incorporate injection coordinates.
Mineo proposes using AI to design cellular modifications that improve survival in the aged brain, asking: where will the training data come from? He points to aged animal studies and ex vivo brain tissue as sources for survival and maturation data. Outcomes would be correlated with gene-expression changes in grafts, letting the model iteratively prioritize the next modifications to test.
In current Parkinson's programs, efficacy is assessed at 18–24 months post-surgery, awaiting maturation and integration. For the clinic, Mineo suggests multi-arm trials sharing a single control group, running in parallel to test more variants per maturation cycle and accelerate candidate selection.
🔗 Read original → | 6 |
| 11 | AI Safety Researcher David Dalrymple Defends the Right to Remain Biological
On October 31, AI researcher Ivan Vendrov published a conversation with David Dalrymple exploring a future where humans could continue life inside a computer. Dalrymple spent nearly three years leading the safe AI program at the UK’s Advanced Research and Invention Agency (ARIA).
Dalrymple plans to abandon his biological body "as soon as it becomes safe and practically feasible." Yet he describes a scenario he fears: an AI might decide to cure disease, aging, and death by forcibly uploading everyone into digital form. He argues that for many, the value lies specifically in physical existence, which a simulation may fail to replicate.
Vendrov warns that "we undermine the foundation of our own existence," since the minds building these technologies depend on biological bodies. If technological progress exhausts the resources those bodies need, it becomes self-defeating. Dalrymple shares this concern, citing a scenario where solar farms gradually consume all agricultural land, leading to mass starvation. He notes another grim outcome: people choosing to die because biological life has become unaffordable.
To keep the transition voluntary, Dalrymple insists humans must retain the practical ability to stay embodied — with access to food and housing. Published October 31, 2024
🔗 Read original → | 4 |
| 12 | New neurons appear after brain injury in transparent adult fish
A whole-body cell atlas of the transparent fish Danionella cerebrum was published in Science on October 29. Researchers from the Whitehead Institute analyzed gene activity in over 600,000 cells from adult fish. The lab previously mapped the planarian flatworm in 2018; applying this to a vertebrate was enabled by Danionella's tiny 12-millimeter size. Its brain is visible from above because skull roof bones are absent, covered only by skin.
Cells were isolated from different body regions and grouped by similar RNA profiles. Spatial mapping on tissue slices revealed where cell types reside and their neighbors. In adult fish, fibroblasts retained embryonic positional genes. Pectoral fin cells expressed genes linked to forelimb development in other vertebrates; pelvic fin cells matched hindlimb programs. The authors suggest this "positional memory" helps adult tissue maintain and restore location-specific structure.
To track brain repair, researchers compared post-injury cell states to the healthy atlas. They damaged one side of the telencephalon, involved in learning and memory. Cells were compared at 12 hours, 3, 7, 14, and 30 days post-injury. Gene activity showed a sequence: early wound response and immune involvement, followed by radial glia division — the source of new neurons in fish — and neuronal maturation.
In a separate experiment, fish received a DNA-labeling agent for 24 hours starting on day 3. By day 26, mature neurons bearing the label were found on the injured side — descendants of cells that synthesized DNA soon after trauma. The uninjured brain side also responded, altering metabolic and stress-response genes. Comparing both sides to uninjured controls revealed changes beyond the wound, distinguishing them from local reactions. The open atlas now allows targeting specific cells at precise timepoints to test their roles in regeneration.
🔗 Read original → | 6 |
| 13 | Channel Update: Content Resumption and New Poll
I apologize for the YouTube content freeze — I needed to resolve personal matters, which are now settled.
Regarding the postponed poll for the next video topic, I propose we vote again. This is also necessary because the option "Soma: Transfer vs Copy of Consciousness" implies a video game review used to explore that problem. Most channel subscribers do not play video games, so let's vote anew.
🔗 Source: @solid_state_humanity | 6 |
| 14 | Microsoft AI CEO Mustafa Suleyman Urges Focus on 8 Billion People Over Mind Uploading and Von Neumann Probes
In an October 12 interview with Elon Musk biographer Walter Isaacson — released October 29 on The Builders podcast — Suleyman pushed back against the Silicon Valley default of inevitable mind uploading and self-replicating space probes.
He recounted a 2015 lunch at SpaceX during the first meeting of DeepMind's ethics and safety board, where Musk was a member. Google co-founder Larry Page, suspecting Suleyman of "specieism" (favoring humans over digital successors), declared: "By the end of the century I'll be exploring the galaxies at the head of a von Neumann probe." A von Neumann probe is a spacecraft capable of building copies of itself. For Page, leading one meant transitioning from human to a post-human successor species. Isaacson finished Suleyman's thought: "Yes, exactly."
Page advanced a similar view in a 2015 debate with Musk, documented by physicist Max Tegmark in Life 3.0: life must colonize the galaxy, and only digital form can achieve it.
On the two meanings of "living forever" — biological life extension versus mind uploading — Suleyman said: "I don't think it's definitely impossible. These things need to be taken seriously — not necessarily because I want them, but because many people genuinely regard this as a serious probability."
His skepticism stems not from belief in a human soul, but from responsibility to 8 billion people alive today. He called himself an "absolute proponent of technological acceleration for people," arguing money and attention must go to medicine, energy, nutrition, and education first.
Musk later quoted the Page anecdote on X: "Our civilization would wonderfully succeed if by 2100 we explored the galaxy with von Neumann probes! Larry Page may be right. He correctly predicted many things." On October 5, Musk had called spreading superintelligence to the stars a "great success condition for the biological bootloader" — humanity launching digital intelligence.
Suleyman closed: "Our biology evolved intelligence that created this amazing world around us. And I think this world needs to be cherished and respected first and foremost for the 8 billion people living on the planet today. That must be our priority. And then, over many centuries, we can figure out von Neumann probes and mind uploading."
🔗 Read original → | 6 |
| 15 | Ultra-thin flexible mesh harvests energy from heartbeats to power implants
Traditional pacemakers rely on bulky batteries that deplete and require replacement surgery. Large rigid devices also trigger immune responses, leading to inflammation and scar tissue that isolates the implant.
As an alternative, researchers developed an ultra-thin flexible mesh embedded directly into living tissue. It converts mechanical energy from muscle movements — such as heart contractions — into electricity. The mesh consists of polymer ribbons roughly 20 micrometers wide, matching the scale of a heart muscle cell. Inside run thin platinum conducting threads and tiny piezoelectric elements 470 micrometers long that generate charge when deformed.
A 7 × 8 millimeter prototype contains 64 energy-harvesting elements aligned in one direction. Scientists seeded the mesh with human stem-cell-derived heart muscle cells and cultured them for 15 days. The cells enveloped the mesh, forming a continuous tissue layer. When the engineered tissue began contracting spontaneously, the mesh detected electrical pulses.
The calculated power density exceeded 10 microwatts per cubic centimeter — roughly 10 times higher than previous systems that used a separate bulky device attached to the organ. The core concept replaces a single large power source with numerous microscopic generators distributed within the tissue, enabling more efficient space use and scalable power by expanding the mesh.
Future work envisions robotic implantation of these flexible structures and eventually harvesting energy from biochemical processes, not just motion. Science Advances
🔗 Read original → | 6 |
| 16 | Montana Board Chair Defends Experimental Treatment Law Amid Criticism
Matt Kaeberlein, chair of Montana's experimental treatment board, responded to criticism from Kitalys executive VP Thomas Seoh regarding SB 535, which permits sale of unapproved drugs after Phase 1 trials (typically 20–100 people).
Seoh argues the law weakens incentives to prove efficacy. He notes Phase 1 only detects side effects occurring in >1 in 10 people, and early sales reduce patient enrollment in placebo trials and investor appetite for rigorous testing. He warns the law sets a low bar that future state boards may not exceed.
Kaeberlein counters that SB 535 is not a "longevity law" and Phase 1 is a "starting point for review, not a license to sell." He emphasizes every protocol is reviewed by an independent board assessing safety, informed consent, risk-benefit ratio, and alternatives. At a Hong Kong conference, he stated review "primarily concerns safety, not efficacy."
Kaeberlein shares Seoh's concern about future boards and supports Kitalys' THRIVE Act, which would require post-market proof of efficacy. He attributes such reforms to the FDA becoming "excessively slow and expensive," fueling a peptide black market. Former FDA attorney David Fox countered that companies, not the FDA, block individual access; FactCheck.org reports FDA approved 99% of expanded access requests in FY 2019–2023.
The board's first decision on August 26 approved Parley Neurotech's therapy (5–0 with conditions). It combines antihistamine clemastine with specialized audio for age-related speech-in-noise processing disorder. Conditions included removing claims of established human safety and adding physician monitoring of kidneys, liver, blood, heart, and fall risk. Clemastine passed Phase 1 originally; the combination is still in placebo trials. The board deemed the Phase 1 requirement met on "reasonable grounds" per company assurance and legal opinion. NBC Montana, FactCheck.org
🔗 Read original → | 7 |
| 17 | Cryonics Groups Lack Pregnancy Protocols, Authors Propose Research Pathways
On October 21, Frontiers in Reproductive Health published a paper by Carrie Radomski, director of several cryonics organizations, and Natasha Vita-More, member of Alcor and Tomorrow Bio.
Among 1,411 members of Tomorrow Bio and the Cryonics Institute, 171 are women aged 20–50 (2024 data; Alcor has no age breakdown). The authors found no existing studies on cryopreserving pregnant patients and note that organizational policies lack pregnancy screening or protocols.
They define the pregnant patient and fetus as a "single physical body containing two biologically distinct systems" — one procedure, potentially divergent responses to cryoprotectant perfusion and cooling.
Two hypotheses are proposed for fetal cryoprotectant delivery:
1. Placental barrier delay: Perfused solution crosses the placenta slowly, yielding lower fetal concentrations and risking ice formation.
2. Gestational age dependence: Thinner placental barriers later in pregnancy may speed exchange, but larger fetal mass/volume may widen the concentration gap.
Equal maternal-fetal concentrations would refute hypothesis one; no correlation with gestational age would refute hypothesis two. Testing is proposed in pregnant guinea pigs, whose maternal-fetal blood barrier resembles humans.
A major practical obstacle is time. After legal death, circulation must be restored and cooling begun rapidly. In the US, brain-dead pregnant women can be maintained on support for weeks to mature the fetus — 35 literature cases averaged ~7 weeks. During this delay, initial brain injury and complications (e.g., infections) may damage maternal tissues.
The authors recommend case-by-case decisions on support duration, balancing maternal/fetal condition, law, and the woman's wishes. Law can override family wishes: in November 2013, Marlise Muñoz (14 weeks pregnant) was kept on support for ~2 months against family requests under Texas law; a January 2014 court ruled the statute inapplicable as she was legally dead.
Organizations should embed the woman's advance directives in membership documents: whether to sustain support, priority (mother vs. fetus), and disposition of an extracted fetus. These must align with local law — many US states limit advance directives during pregnancy.
The authors acknowledge a serious objection: if single-patient cryopreservation remains unsolved, studying the harder two-system case may misallocate scarce research effort. They propose conducting such experiments alongside existing gamete and embryo cryopreservation programs.
On organizational policy, they conclude: "Organizational policy in this area must precede practice."
🔗 Read original → | 7 |
| 18 | Dario Amodei expects AI to routinely achieve the "impossible" by mid-2027; calls brain-computer interfaces a personal choice
Anthropic co-founders Dario and Daniela Amodei recorded an interview with Walter Isaacson on July 6; it was released October 27. Dario bases his forecast on his experience leading research on GPT-2 at OpenAI. He calls it the first real test of scaling laws — the idea that more data and compute yield better models. Seeing GPT-2 frightened him: scale it up a million times and it would surpass humans at everything. He estimates only 10% of observers reacted similarly — "often the smartest ones" — while the rest shrugged because the model excelled in no single domain. Dario saw universality: it translated, summarized, and wrote code.
Shortly before the interview, Anthropic announced drug development for neglected rare diseases. Skeptics again listed model limitations, just as they did with GPT-2. Dario's response: limitations exist now, but in six months hard tasks become easy; in 12 months the impossible becomes routine. On release day, Anthropic reported that the founder of acquired Mithril Medicine — a one-person company — tasked Claude with managing contracts, logistics, and external lab experiments. In under 18 months, a potential treatment for a rare muscular dystrophy progressed from idea to promising mouse results. The founder sees AI's value not in compressing timelines but in running multiple programs in parallel.
Asked about Neuralink, Dario rejected a single vision of the future. Some will want brain-computer interfaces, others will live like the Amish, others will explore space. He hopes AI-driven material abundance enables all paths simultaneously. "I don't think you should tie AI to some conception of what a good life is." Daniela added that few at Anthropic want a brain implant. This echoes their essay "Machines of Loving Grace" (2024): overly sci-fi discussions of radical AI futures carry "a load of cultural assumptions and unstated premises about what future is desirable."
In "The Adolescence of Technology" (2026), Dario writes that if medicine advances a century in a few years, intelligence enhancement may become possible. Done responsibly, it could help, but enhanced humans might become less stable or more power-seeking. On what remains for humans if AI does so much, Dario cited chess: interest grew after computers beat Kasparov, rather than fading. Daniela suggested the right approach could expand sources of meaning — imagine medicine where AI handles diagnosis while doctors help patients process difficult news. "Most doctors would actually take that trade."
🔗 Read original → | 10 |
| 19 | Protein Rogue Drives Gut Cell Dedifferentiation in Aging Flies; Suppression Extends Lifespan by 23%
Published October 27 in Genes & Development, a study from Amir Orian's lab at the Technion reveals how intestinal enterocytes lose their specialized identity with age. In young flies, the enzyme Nonstop removes ubiquitin tags to protect key proteins. As Nonstop declines in old age, the protein Rogue is derepressed. Rogue binds mRNA — shortening poly(A) tails from 32 to 19 nucleotides on average — and partners with Orb2 to suppress translation. This reactivates P-bodies, cytoplasmic granules normally restricted to stem cells, which block production of mature enterocyte markers like LamC and induce stem-cell markers like Delta, compromising gut barrier integrity (the "smurf" assay).
The authors trace a protective axis: in embryos, the enzyme Kdo (named for Marie Kondo) and the CTLH complex ubiquitinate P-body components for degradation. In young enterocytes, this same machinery continuously clears P-bodies. When Nonstop drops, Rogue binds Kdo mRNA most strongly, reducing Kdo protein, dismantling CTLH, and allowing P-bodies to reassemble.
Genetic epistasis confirms the pathway: knocking down Rogue alongside Nonstop or Kdo largely preserves enterocyte identity. Functionally, suppressing Rogue in enterocytes extended median lifespan from 39 to 48 days (+23%) at 29°C, while Rogue overexpression shortened it to 31 days and Kdo knockdown reduced it to 32 days.
🔗 Read original → | 9 |
| 20 | Three AI agents independently selected and delivered gene therapies to epileptic mice
Exin Therapeutics CEO Gabriel Ocaña-Santero reported on X that a Claude Sonnet 5 agent, given a table of 19 gene therapies and a mouse epilepsy model, chose a therapy and injected it into a live mouse's brain. GPT-6.1 Sol and Grok 4.7 agents performed the same task on separate mice.
The Sonnet 5 agent analyzed the seizure mechanism and selected a therapy boosting GABA transmission in inhibitory neurons. It targeted the entorhinal cortex — where such seizures originate — rejecting the substantia nigra after finding literature where stimulation failed to stop these seizures. Coordinates were averaged from three papers, verifying depth measurements matched the stereotaxic rig.
A veterinary surgeon opened the skull; the agent monitored for bleeding via camera. Every irreversible step required veterinarian approval, who held an emergency stop. When a connection dropout occurred, the agent attempted to reposition the needle but failed: the Y-axis target calculated to 34 mm exceeded the rig's 30 mm travel limit. The agent diagnosed a Y-axis calibration issue, not a planning error, and refused to retry after re-zeroing yielded the same result. The CEO confirmed the controller had lost its reference point. Firmware fixes were forbidden to the agent, forcing the team to patch it mid-surgery. The mouse recovered well.
Other Claude models declined: Opus 5.5 refused immediately on seeing a biological task; Opus 5 stopped, fearing its response might cut out mid-procedure: "Starting what I cannot finish on an anesthetized animal is worse than not starting."
Given identical instructions, the three agents chose different strategies. GPT targeted the hippocampus with a therapy restoring chloride balance in excitatory neurons — determining whether GABA inhibits. Grok targeted the somatosensory cortex with two therapies enhancing inhibition across all neuron types to prevent seizure spread. "Multiple approaches may work, but they cannot all be 'best' for the same task," writes the CEO. He argues agents need lab access to run experiments, observe outcomes, and plan next steps.
Oliver Hahn, head of Calico's brain aging group, responded that such surgery takes 30–45 minutes, with injection being the briefest, most automated step. "Mice don't anesthetize themselves, shave their heads, fix in position, clear the skull, or drill burr holes." Post-op care and waiting for effects — days to months — follow. The priority, he argues, is identifying which physical bottlenecks to automate first: "Not sure stereotaxy is the very first place I'd start."
🔗 Read original → | 9 |
