PromptBio releases preprint on PromptGenie system for biomedical computation tracking
On September 6, the PromptBio team posted a preprint on bioRxiv describing PromptGenie, a system for biomedical data research. The authors explain how it turns a research question into a plan, code, execution, and records that allow verification of the computational workflow. A biomedical question is transformed into output through a series of decisions: what data to use, which method to test a hypothesis, how to perform the calculation, and how to interpret results.
When these decisions reside in separate files, services, and email threads, it is hard to see from a final table or figure which data and choices led to it. In the preprint, the authors call PromptGenie a system that receives a research goal and available context, builds a plan of dependent steps, and adjusts subsequent steps based on intermediate results or errors. Before a major decision, the researcher can review and edit the plan.
The system then invokes a ready-made method or writes code, runs the calculation, and passes the result to the next step. Authors describe a journal where the plan and its modifications, data sources and references, parameters, execution environment details,
🔗 Read original →
On September 6, the PromptBio team posted a preprint on bioRxiv describing PromptGenie, a system for biomedical data research. The authors explain how it turns a research question into a plan, code, execution, and records that allow verification of the computational workflow. A biomedical question is transformed into output through a series of decisions: what data to use, which method to test a hypothesis, how to perform the calculation, and how to interpret results.
When these decisions reside in separate files, services, and email threads, it is hard to see from a final table or figure which data and choices led to it. In the preprint, the authors call PromptGenie a system that receives a research goal and available context, builds a plan of dependent steps, and adjusts subsequent steps based on intermediate results or errors. Before a major decision, the researcher can review and edit the plan.
The system then invokes a ready-made method or writes code, runs the calculation, and passes the result to the next step. Authors describe a journal where the plan and its modifications, data sources and references, parameters, execution environment details,
🔗 Read original →
Suppressing MEOX1 improves cardiac reprogramming after heart attack in mice
On September 8, researchers published in Circulation a study in which they treated mice after myocardial infarction with a cocktail of cardiac‑reprogramming factors together with a short RNA designed to suppress the MEOX1 gene.
Six weeks later, the mice that received MEOX1‑suppressing RNA had more cells expressing cardiac‑muscle markers, a smaller scar, and a markedly improved ejection fraction (38.1% versus 28.85% in the control group). The number of spontaneously contracting cells was also increased roughly about fourfold.
Further analysis showed that MEOX1 drives a scar‑promoting program and is upregulated by the profibrotic signal TGF‑β1; reducing its activity allowed fibroblasts to adopt a cardiomyocyte fate, whereas restoring MEOX1 or using a RNA‑resistant version abolished the benefit.
The same combination of reprogramming factors and MEOX1 inhibition also raised the proportion of cardiac‑muscle‑positive cells in cultured human fibroblasts, supporting the idea that blocking MEOX1 helps convert scar‑forming cells into functional heart muscle.
🔗 Read original →
On September 8, researchers published in Circulation a study in which they treated mice after myocardial infarction with a cocktail of cardiac‑reprogramming factors together with a short RNA designed to suppress the MEOX1 gene.
Six weeks later, the mice that received MEOX1‑suppressing RNA had more cells expressing cardiac‑muscle markers, a smaller scar, and a markedly improved ejection fraction (38.1% versus 28.85% in the control group). The number of spontaneously contracting cells was also increased roughly about fourfold.
Further analysis showed that MEOX1 drives a scar‑promoting program and is upregulated by the profibrotic signal TGF‑β1; reducing its activity allowed fibroblasts to adopt a cardiomyocyte fate, whereas restoring MEOX1 or using a RNA‑resistant version abolished the benefit.
The same combination of reprogramming factors and MEOX1 inhibition also raised the proportion of cardiac‑muscle‑positive cells in cultured human fibroblasts, supporting the idea that blocking MEOX1 helps convert scar‑forming cells into functional heart muscle.
🔗 Read original →
PubMed Central (PMC)
In Vivo Targeted Reprogramming of Cardiac Fibroblasts for Heart Regeneration: Advances and Therapeutic Potential
Myocardial infarction-induced cardiovascular diseases remain a leading cause of mortality worldwide. Excessive post-infarct fibrosis contributes to adverse cardiac remodeling and the progression to heart failure. In vivo reprogramming strategies ...
FDA talks preventive therapies for multiple age-related diseases at H-SPAN summit
On September 9, the H‑SPAN longevity biotech summit hosted a panel on how the FDA should evaluate preventive therapies aimed at several age‑related diseases. The discussion built on questions raised in the July H‑SPAN summit report about which disease to target first and what metric to use to prove benefit.
Robert Kalesnik-Orshulak, who leads regulatory affairs at Retro Biosciences, argued that companies should start with a single indication, obtain early results, and then expand the label to other aging‑related conditions. He suggested using the FDA’s accelerated approval pathway, where a surrogate endpoint could serve as the basis for a decision.
David Fox, an FDA regulatory lawyer and co‑author of the Thrive Act, countered that testing a therapy meant to delay multiple chronic diseases before a serious diagnosis requires large cohorts, long follow‑up, and especially rigorous safety assessment. He noted that his Thrive Act outlines a staged evidence framework for prevention in people without a severe diagnosis.
Lowell Zeta, introduced as the FDA deputy commissioner for strategy and acting chief of staff, invited industry to submit data and proposals. He pointed to **Operation
🔗 Read original →
On September 9, the H‑SPAN longevity biotech summit hosted a panel on how the FDA should evaluate preventive therapies aimed at several age‑related diseases. The discussion built on questions raised in the July H‑SPAN summit report about which disease to target first and what metric to use to prove benefit.
Robert Kalesnik-Orshulak, who leads regulatory affairs at Retro Biosciences, argued that companies should start with a single indication, obtain early results, and then expand the label to other aging‑related conditions. He suggested using the FDA’s accelerated approval pathway, where a surrogate endpoint could serve as the basis for a decision.
David Fox, an FDA regulatory lawyer and co‑author of the Thrive Act, countered that testing a therapy meant to delay multiple chronic diseases before a serious diagnosis requires large cohorts, long follow‑up, and especially rigorous safety assessment. He noted that his Thrive Act outlines a staged evidence framework for prevention in people without a severe diagnosis.
Lowell Zeta, introduced as the FDA deputy commissioner for strategy and acting chief of staff, invited industry to submit data and proposals. He pointed to **Operation
🔗 Read original →
U.S. Food and Drug Administration
Accelerated Approval Program
Drug and Biologic Accelerated Approvals Based on a Surrogate Endpoint
Experts propose labeling aging a 'metadisease' to streamline clinical trials
At the H-SPAN summit on 9 September, anatomist Antonney Chok urged that aging be described as a “metadisease” — a biological process that underlies several age‑related conditions. He argued that this framing would help clinical trials assess interventions targeting aging biology by measuring a set of related outcomes.
The discussion built on the July report of the H-SPAN summit, which devoted its second day to clinical endpoints and FDA regulations. Panelists asked what a single trial should test if a therapy alters the overall aging process, noting that endpoints must be predefined and measurable.
Chok explained that because aging drives multiple pathologies, an intervention will need a bundle of linked metrics rather than a single disease‑specific endpoint. He highlighted cellular senescence — the state in which cells cease dividing and alter tissue function — as a central mechanism in his model.
To enable one trial to evaluate several results, Chok proposed assigning an ICD code to aging as a metadisease. He also suggested creating a federal institute of gerontology alongside the National Institute on Aging to unite researchers who currently work in cancer, heart disease, or Alzheimer’s programs around shared aging biology.
🔗 Read original →
At the H-SPAN summit on 9 September, anatomist Antonney Chok urged that aging be described as a “metadisease” — a biological process that underlies several age‑related conditions. He argued that this framing would help clinical trials assess interventions targeting aging biology by measuring a set of related outcomes.
The discussion built on the July report of the H-SPAN summit, which devoted its second day to clinical endpoints and FDA regulations. Panelists asked what a single trial should test if a therapy alters the overall aging process, noting that endpoints must be predefined and measurable.
Chok explained that because aging drives multiple pathologies, an intervention will need a bundle of linked metrics rather than a single disease‑specific endpoint. He highlighted cellular senescence — the state in which cells cease dividing and alter tissue function — as a central mechanism in his model.
To enable one trial to evaluate several results, Chok proposed assigning an ICD code to aging as a metadisease. He also suggested creating a federal institute of gerontology alongside the National Institute on Aging to unite researchers who currently work in cancer, heart disease, or Alzheimer’s programs around shared aging biology.
🔗 Read original →
World Health Organization
Ageing and health
Key facts on ageing and health: explanation of ageing, common health conditions associated with ageing, Factors influencing Healthy Ageing, Challenges in responding to population ageing and WHO's response
Jay Olshansky Proposes Measuring Medical Success by Years of Independent Living
At the H-SPAN meeting on the economics of aging on September 9, aging researcher Jay Olshansky called the later years of life a 'red zone' where the risk of frailty and disability rises rapidly. He proposed measuring medical progress by how many years of independent living can be preserved.
In the XX век, public health and medicine shifted most deaths from younger to older ages. Olshansky’s next step is to delay the period when frailty, disability, and disease accumulate together with age.
In his model, the 'red zone' is the late-life period with rapidly increasing risk of these conditions. Treating a single diagnosis may avert one threat but leaves time to encounter other age‑related damages, as he summarized with the formula: 'Knock down one disease — two rise; knock down two — four.'
By targeting the processes that cause damage to accumulate with age, onset of multiple age‑related impairments could be postponed and frailty years reduced. Success is measured in years during which a person maintains mobility, clear thinking, and the ability to live without constant assistance.
The
🔗 Read original →
At the H-SPAN meeting on the economics of aging on September 9, aging researcher Jay Olshansky called the later years of life a 'red zone' where the risk of frailty and disability rises rapidly. He proposed measuring medical progress by how many years of independent living can be preserved.
In the XX век, public health and medicine shifted most deaths from younger to older ages. Olshansky’s next step is to delay the period when frailty, disability, and disease accumulate together with age.
In his model, the 'red zone' is the late-life period with rapidly increasing risk of these conditions. Treating a single diagnosis may avert one threat but leaves time to encounter other age‑related damages, as he summarized with the formula: 'Knock down one disease — two rise; knock down two — four.'
By targeting the processes that cause damage to accumulate with age, onset of multiple age‑related impairments could be postponed and frailty years reduced. Success is measured in years during which a person maintains mobility, clear thinking, and the ability to live without constant assistance.
The
🔗 Read original →
Nature
Implausibility of radical life extension in humans in the twenty-first century
Nature Aging - In the twentieth century, human life expectancy rose dramatically. Based on the past three decades of observed mortality in the eight countries with the longest-lived populations and...
H-SPAN Summit Links NIH Grants, Lab Timing, and Payer Decisions
At the H-SPAN summit on longevity policy held on September 9, Felipe Sierra, Catherine Kachorovski, and Michael Rinkel discussed how grants, clinical data, and payer decisions form a chain. The US National Institutes of Health (NIH) allocate grants, labs depend on their timing, and insurers or government programs decide whether to pay for treatment based on trial data.
Felipe Sierra, former director of the National Institute on Aging’s Biology of Aging division, explained why his division lacked funding: the number of applications grew while the budget stayed flat. He said he did not ask for more money but wanted NIH to spend its funds elsewhere—specifically on aging‑as‑a‑common‑risk‑factor research already pursued by other NIH institutes. He urged their leaders to redirect their own budgets toward such projects.
Catherine Kachorovski, a professor at the University of Michigan, showed how a grant delay affects a lab. Her application scored in the top 5%; she was initially told she would be funded, but the money was held up because part of the budget had already been pledged to five‑year grants. Universities cut graduate‑student admissions, labs lost funding, and for her lab the delay influenced which staff she could retain.
Michael Rinkel, COO of biotech firm Life Biosciences, shifted the talk to drugs. Phase I trials give early safety and dosage data; later studies test possible efficacy and gather more safety information. He noted that each year 200–300 drug candidates pass Phase I but fail Phase II, so he opposes releasing a drug after only early safety checks.
In the spring the New Hampshire House of Representatives approved centers where drugs can be used after Phase I. This makes Rinkel’s debate practical: payers can decide reimbursement based on data collected after early safety testing. He also warned that lengthy validation is costly, and an analysis of cheap gene‑therapy trials showed why a therapy that is easy to copy may attract no investor for such study—its benefits would be available to all providers.
🔗 Read original →
At the H-SPAN summit on longevity policy held on September 9, Felipe Sierra, Catherine Kachorovski, and Michael Rinkel discussed how grants, clinical data, and payer decisions form a chain. The US National Institutes of Health (NIH) allocate grants, labs depend on their timing, and insurers or government programs decide whether to pay for treatment based on trial data.
Felipe Sierra, former director of the National Institute on Aging’s Biology of Aging division, explained why his division lacked funding: the number of applications grew while the budget stayed flat. He said he did not ask for more money but wanted NIH to spend its funds elsewhere—specifically on aging‑as‑a‑common‑risk‑factor research already pursued by other NIH institutes. He urged their leaders to redirect their own budgets toward such projects.
Catherine Kachorovski, a professor at the University of Michigan, showed how a grant delay affects a lab. Her application scored in the top 5%; she was initially told she would be funded, but the money was held up because part of the budget had already been pledged to five‑year grants. Universities cut graduate‑student admissions, labs lost funding, and for her lab the delay influenced which staff she could retain.
Michael Rinkel, COO of biotech firm Life Biosciences, shifted the talk to drugs. Phase I trials give early safety and dosage data; later studies test possible efficacy and gather more safety information. He noted that each year 200–300 drug candidates pass Phase I but fail Phase II, so he opposes releasing a drug after only early safety checks.
In the spring the New Hampshire House of Representatives approved centers where drugs can be used after Phase I. This makes Rinkel’s debate practical: payers can decide reimbursement based on data collected after early safety testing. He also warned that lengthy validation is costly, and an analysis of cheap gene‑therapy trials showed why a therapy that is easy to copy may attract no investor for such study—its benefits would be available to all providers.
🔗 Read original →
U.S. Food and Drug Administration
Step 3: Clinical Research
Step 3: Clinical Research
Single IV injection of dual‑gene AAV9 extends lifespan in Sandhoff cat model
In a study published Science Translational Medicine, September 9, researchers used a single intravenous injection of an AAV9 vector carrying both HEXA and HEXB genes to treat asymptomatic kittens with a feline model of Sandhoff disease. The vector aims to restore the HexA enzyme needed to break down the lipid GM2 that accumulates in lysosomes.
Untreated animals survived a median of 4.3 months, while low‑dose kittens lived 8.3 months on average. High‑dose treatment extended median survival to 12.4 months, and one cat reached 16 months.
Tremor appeared in untreated kittens around 2.4 months of age. No tremor was observed in the high‑dose group, whereas tremor was delayed in three of four low‑dose kittens.
Sixteen weeks after the high‑dose injection, MRI showed partial preservation of brain structure and CSF levels of neurodegeneration markers were reduced. HexA activity reached approximately 70% of normal in the hindbrain and 120% in the spinal cord.
The liver contained the highest vector copy number, but HexA activity there did not exceed 9% of normal levels. Over longer periods, brain architecture and inflammatory signs in some treated animals began to resemble those of untreated controls.
🔗 Read original →
In a study published Science Translational Medicine, September 9, researchers used a single intravenous injection of an AAV9 vector carrying both HEXA and HEXB genes to treat asymptomatic kittens with a feline model of Sandhoff disease. The vector aims to restore the HexA enzyme needed to break down the lipid GM2 that accumulates in lysosomes.
Untreated animals survived a median of 4.3 months, while low‑dose kittens lived 8.3 months on average. High‑dose treatment extended median survival to 12.4 months, and one cat reached 16 months.
Tremor appeared in untreated kittens around 2.4 months of age. No tremor was observed in the high‑dose group, whereas tremor was delayed in three of four low‑dose kittens.
Sixteen weeks after the high‑dose injection, MRI showed partial preservation of brain structure and CSF levels of neurodegeneration markers were reduced. HexA activity reached approximately 70% of normal in the hindbrain and 120% in the spinal cord.
The liver contained the highest vector copy number, but HexA activity there did not exceed 9% of normal levels. Over longer periods, brain architecture and inflammatory signs in some treated animals began to resemble those of untreated controls.
🔗 Read original →
Review of 101 Species Highlights DNA Maintenance, Tissue Renewal, Stress Resistance
On 8 September the journal Aging published a review of 101 species with maximum lifespans starting at 250 years, comprising 11 animals and 90 plants.
The authors compared lifespan differences among individuals of the same species and between different species, asking why individuals of a long‑lived species vary in longevity and how these species differ from shorter‑lived ones; they sought processes that recur in both comparisons.
In both comparisons three common directions emerged: DNA maintenance (repairing damage), stem cell properties (enabling cells to divide and generate different tissue types), and stress resistance (helping cells cope with unfavorable conditions and associated harm).
When comparing species, the authors also examined regeneration and tumor suppression; in hydra regeneration replaces all body cells roughly every twenty days, while in common pine and ginkgo root‑growth zone cells retain differentiation ability with age, unlike the shorter‑lived Arabidopsis where this ability declines.
The authors link the three common directions to early developmental stages and germ cells, concluding with the phrase: “Staying young, it seems, is a strategy for long life.”
🔗 Read original →
On 8 September the journal Aging published a review of 101 species with maximum lifespans starting at 250 years, comprising 11 animals and 90 plants.
The authors compared lifespan differences among individuals of the same species and between different species, asking why individuals of a long‑lived species vary in longevity and how these species differ from shorter‑lived ones; they sought processes that recur in both comparisons.
In both comparisons three common directions emerged: DNA maintenance (repairing damage), stem cell properties (enabling cells to divide and generate different tissue types), and stress resistance (helping cells cope with unfavorable conditions and associated harm).
When comparing species, the authors also examined regeneration and tumor suppression; in hydra regeneration replaces all body cells roughly every twenty days, while in common pine and ginkgo root‑growth zone cells retain differentiation ability with age, unlike the shorter‑lived Arabidopsis where this ability declines.
The authors link the three common directions to early developmental stages and germ cells, concluding with the phrase: “Staying young, it seems, is a strategy for long life.”
🔗 Read original →
Aging-Us
How to live for centuries: common denominators of organisms with exceptional longevity | Aging
Organisms vary in their lifespan. Understanding this variation may help us live healthier and longer. Here, we focus on species living twice as long as humans, or more. Out of the 101 multicellular s
Southern Cryonics Stores Seven Bodies in Liquid Nitrogen at Holbrook Facility
On September 10, a Guardian reporter visited the Southern Cryonics facility in Holbrook, New South Wales and observed seven bodies stored in two liquid‑nitrogen dewars. The dewars keep the bodies upside down at -196 °C, and the supplier regularly tops off the tanks to prevent dangerous pressure buildup from evaporating nitrogen.
At the request of its first clients, Southern Cryonics shortened the delivery interval from one month to two‑week cycles. Preparation for storage begins when a patient reaches a terminal state, with CryoPath verifying funding, assembling a standby team and arranging a funeral home.
After legal death, a perfusionist cools the body and adds a cryoprotectant to inhibit ice‑crystal formation; the body is then shipped to Holbrook on dry ice. With direct freezing, the funeral home handles stabilization.
Cryobiologist Saffron Bryant of the Royal Melbourne Institute of Technology notes that vitrification already preserves cells, embryos and blood products, but delivering cryoprotectant evenly throughout whole organs or bodies remains unsolved.
The published Southern Cryonics contract states that continued storage depends on the client organizing and paying for standby, stabilization and transport; failures such as lack of funds, unpaid fees, delays, loss of permit, legal changes or operator insolvency allow the company to refuse or halt cryopreservation, after which the body may be buried, cremated or transferred to the legal custodian.
🔗 Read original →
On September 10, a Guardian reporter visited the Southern Cryonics facility in Holbrook, New South Wales and observed seven bodies stored in two liquid‑nitrogen dewars. The dewars keep the bodies upside down at -196 °C, and the supplier regularly tops off the tanks to prevent dangerous pressure buildup from evaporating nitrogen.
At the request of its first clients, Southern Cryonics shortened the delivery interval from one month to two‑week cycles. Preparation for storage begins when a patient reaches a terminal state, with CryoPath verifying funding, assembling a standby team and arranging a funeral home.
After legal death, a perfusionist cools the body and adds a cryoprotectant to inhibit ice‑crystal formation; the body is then shipped to Holbrook on dry ice. With direct freezing, the funeral home handles stabilization.
Cryobiologist Saffron Bryant of the Royal Melbourne Institute of Technology notes that vitrification already preserves cells, embryos and blood products, but delivering cryoprotectant evenly throughout whole organs or bodies remains unsolved.
The published Southern Cryonics contract states that continued storage depends on the client organizing and paying for standby, stabilization and transport; failures such as lack of funds, unpaid fees, delays, loss of permit, legal changes or operator insolvency allow the company to refuse or halt cryopreservation, after which the body may be buried, cremated or transferred to the legal custodian.
🔗 Read original →
the Guardian
‘Patients’ who pay $200,000 to be frozen after death are waiting to be reborn in an Australian shed
Exclusive: Southern Cryonics operates out of a shed in Holbrook in rural NSW. Experts say reanimation is ‘a dream’ but true believers are willing to pay for a second chance at life
Old fibroblasts transfer polarity defect to young cells via ApoD
The bioRxiv, September 10 preprint describes a loop in which ApoD secreted by aged fibroblasts raises SUN1 at the nuclear envelope, disrupting the orientation of young fibroblasts. High SUN1 levels in turn increase ApoD secretion, establishing a positive feedback loop.
Cell polarity requires fibroblasts to align a front‑rear axis, positioning the nucleus and the microtubule‑organizing center. When young fibroblasts were cultured in medium previously conditioned by old cells, they failed to correctly place the nucleus and center after a motility cue, whereas control medium had no effect. Proteomic analysis of the conditioned medium implicated ApoD, and purified ApoD reproduced the polarity defect.
ApoD elevated SUN1 in young fibroblasts, strengthening the SUN1–nesprin‑2 bridge that links the nucleus to microtubules; excess SUN1 blocked proper nuclear positioning. Knocking down SUN1 or disrupting the microtubule‑nucleus tether rescued polarity, and high SUN1 also drove more ApoD release, while SUN1 suppression lowered ApoD secretion.
The defect extended to muscle progenitors: ApoD impaired their ability to form fibers in culture. In vivo, mice received ten days of daily intravenous ApoD injections; on day‑five half underwent muscle injury. Injured muscle showed reduced mass and contractile force versus controls, while uninjured muscle was unchanged.
🔗 Read original →
The bioRxiv, September 10 preprint describes a loop in which ApoD secreted by aged fibroblasts raises SUN1 at the nuclear envelope, disrupting the orientation of young fibroblasts. High SUN1 levels in turn increase ApoD secretion, establishing a positive feedback loop.
Cell polarity requires fibroblasts to align a front‑rear axis, positioning the nucleus and the microtubule‑organizing center. When young fibroblasts were cultured in medium previously conditioned by old cells, they failed to correctly place the nucleus and center after a motility cue, whereas control medium had no effect. Proteomic analysis of the conditioned medium implicated ApoD, and purified ApoD reproduced the polarity defect.
ApoD elevated SUN1 in young fibroblasts, strengthening the SUN1–nesprin‑2 bridge that links the nucleus to microtubules; excess SUN1 blocked proper nuclear positioning. Knocking down SUN1 or disrupting the microtubule‑nucleus tether rescued polarity, and high SUN1 also drove more ApoD release, while SUN1 suppression lowered ApoD secretion.
The defect extended to muscle progenitors: ApoD impaired their ability to form fibers in culture. In vivo, mice received ten days of daily intravenous ApoD injections; on day‑five half underwent muscle injury. Injured muscle showed reduced mass and contractile force versus controls, while uninjured muscle was unchanged.
🔗 Read original →
Astra AI designs complex robotic arm and paints Golden Gate Bridge in real time
Astra generated 190 ready‑to‑use CAD files in STEP format for assembling a highly complex robotic arm, using just 48 hours of work and ~50,000 lines of Python code. The AI didn’t only draw the geometry; it fully programmed the kinematics, visualized the result with JavaScript, and exported the finished 3D models as GLB files for rendering.
In the second video, Astra was given a physical manipulator, a paint‑filled brush, and a camera for feedback, then tasked with painting the Golden Gate Bridge. The model independently worked out the robot’s kinematics, learned to evaluate the output via the camera, and improved each successive stroke through real‑time trial and error until it produced a coherent painting.
🔗 Source: @solid_state_humanity
Astra generated 190 ready‑to‑use CAD files in STEP format for assembling a highly complex robotic arm, using just 48 hours of work and ~50,000 lines of Python code. The AI didn’t only draw the geometry; it fully programmed the kinematics, visualized the result with JavaScript, and exported the finished 3D models as GLB files for rendering.
In the second video, Astra was given a physical manipulator, a paint‑filled brush, and a camera for feedback, then tasked with painting the Golden Gate Bridge. The model independently worked out the robot’s kinematics, learned to evaluate the output via the camera, and improved each successive stroke through real‑time trial and error until it produced a coherent painting.
🔗 Source: @solid_state_humanity
Telegram
Solid State Humanity
Многообещающие достижения Astra
1-ое видео - всего за 48 часов работы и ~50 000 строк кода на Python Astra сгенерировала 190 готовых CAD-файлов (формат STEP) для сборки сложнейшей роборуки. ИИ не просто нарисовал геометрию, а полностью запрограммировал кинематику…
1-ое видео - всего за 48 часов работы и ~50 000 строк кода на Python Astra сгенерировала 190 готовых CAD-файлов (формат STEP) для сборки сложнейшей роборуки. ИИ не просто нарисовал геометрию, а полностью запрограммировал кинематику…