#materialsscience — Public Fediverse posts

What Has AI Actually Solved? The Biggest Problems Where Machines Found Real Answers

Dear Cherubs, AI has been promised as everything from an office assistant to the machine that will apparently fix civilisation before lunch. Strip away the hype, though, and something genuinely remarkable has happened: AI has already helped solve, or substantially crack, several problems that humans had struggled with for decades.

The important bit is “helped”. AI has not cured cancer, solved climate change or discovered the meaning of life while we were making coffee. But in science, mathematics, medicine and engineering, it has produced results that have survived contact with reality.

PROBLEMS THAT GOT REAL ANSWERS

Perhaps the most famous example is protein structure. For decades, predicting how a protein folds into its three-dimensional shape was a major biological challenge. DeepMind’s AlphaFold transformed the field by predicting protein structures at unprecedented scale, contributing to work that earned Demis Hassabis and John Jumper the 2024 Nobel Prize in Chemistry.

Then there are antibiotics. MIT researchers used machine learning to identify halicin, a previously known compound with powerful antibacterial activity, including against several drug-resistant bacteria. Later work identified abaucin, a new antibiotic candidate targeting Acinetobacter baumannii, with activity demonstrated in laboratory experiments and a mouse infection model. MIT’s Antibiotics-AI project now reports nine novel antibiotics discovered or designed with AI and more than 70 billion molecules screened computationally.

Materials science has produced another eye-watering number. Google DeepMind’s GNoME system predicted around 2.2 million crystal structures, including 380,000 predicted to be particularly stable. Researchers independently produced 736 of those predicted structures experimentally, turning some computer-generated possibilities into actual materials.

Mathematics has also taken a hit from the robot army. In 2024, DeepMind reported that AlphaProof and AlphaGeometry solved four of six International Mathematical Olympiad problems, reaching silver-medal-level performance. AlphaProof also produced formal proofs for its solutions, rather than merely throwing suspiciously confident numbers at the wall.

FROM ANSWERS TO DISCOVERY

The more interesting development may be that AI is increasingly being used not simply to answer questions, but to generate new solutions.

AlphaEvolve, introduced by Google DeepMind in 2025, combines large language models with automated evaluation and evolutionary search to discover and improve algorithms. Reported applications include faster matrix multiplication, improvements to computing infrastructure and solutions to mathematical problems. By 2026, Google reported further work involving DNA sequencing, disaster prediction, power-grid simulations and scientific computing.

That changes the game slightly. Traditional science usually looks something like: human has idea, human tests idea, human gets annoyed, human tries another idea.

AI can potentially run the cycle at enormous scale: generate thousands of hypotheses, test them computationally, discard the rubbish and send the promising candidates to humans or automated laboratories.

And that is probably the real story.

AI has not “solved everything”. It has started solving pieces of problems that once seemed stubbornly resistant to computation, while opening entirely new ways of searching through possibilities.

The next breakthrough may therefore not be one spectacular answer.

It may be the machine that keeps finding the next question.

Sources list:

Nobel Prize — The Nobel Prize in Chemistry 2024 — https://www.nobelprize.org/prizes/chemistry/2024/summary/

Google DeepMind — Millions of new materials discovered with deep learning — https://deepmind.google/blog/millions-of-new-materials-discovered-with-deep-learning/

Google DeepMind — AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms — https://deepmind.google/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/

Google DeepMind — AI achieves silver-medal standard solving International Mathematical Olympiad problems — https://deepmind.google/blog/ai-solves-imo-problems-at-silver-medal-level/

Google — AlphaEvolve, 1 year later: Impact on science, technology — https://blog.google/innovation-and-ai/infrastructure-and-cloud/google-cloud/alphaevolve-updates/

MIT News — Artificial intelligence yields new antibiotic — https://news.mit.edu/2020/artificial-intelligence-identifies-new-antibiotic-0220

MIT Jameel Clinic — Antibiotic identified by AI — https://jclinic.mit.edu/antibiotic-identified-by-ai/

MIT Collins Lab — Antibiotics-AI Project — https://www.collinslab.mit.edu/antibiotics-ai-project

Thisclaimer — https://thisclaimer.com

What Has AI Actually Solved? The Biggest Problems Where Machines Found Real Answers

Dear Cherubs, AI has been promised as everything from an office assistant to the machine that will apparently fix civilisation before lunch. Strip away the hype, though, and something genuinely remarkable has happened: AI has already helped solve, or substantially crack, several problems that humans had struggled with for decades.

The important bit is “helped”. AI has not cured cancer, solved climate change or discovered the meaning of life while we were making coffee. But in science, mathematics, medicine and engineering, it has produced results that have survived contact with reality.

PROBLEMS THAT GOT REAL ANSWERS

Perhaps the most famous example is protein structure. For decades, predicting how a protein folds into its three-dimensional shape was a major biological challenge. DeepMind’s AlphaFold transformed the field by predicting protein structures at unprecedented scale, contributing to work that earned Demis Hassabis and John Jumper the 2024 Nobel Prize in Chemistry.

Then there are antibiotics. MIT researchers used machine learning to identify halicin, a previously known compound with powerful antibacterial activity, including against several drug-resistant bacteria. Later work identified abaucin, a new antibiotic candidate targeting Acinetobacter baumannii, with activity demonstrated in laboratory experiments and a mouse infection model. MIT’s Antibiotics-AI project now reports nine novel antibiotics discovered or designed with AI and more than 70 billion molecules screened computationally.

Materials science has produced another eye-watering number. Google DeepMind’s GNoME system predicted around 2.2 million crystal structures, including 380,000 predicted to be particularly stable. Researchers independently produced 736 of those predicted structures experimentally, turning some computer-generated possibilities into actual materials.

Mathematics has also taken a hit from the robot army. In 2024, DeepMind reported that AlphaProof and AlphaGeometry solved four of six International Mathematical Olympiad problems, reaching silver-medal-level performance. AlphaProof also produced formal proofs for its solutions, rather than merely throwing suspiciously confident numbers at the wall.

FROM ANSWERS TO DISCOVERY

The more interesting development may be that AI is increasingly being used not simply to answer questions, but to generate new solutions.

AlphaEvolve, introduced by Google DeepMind in 2025, combines large language models with automated evaluation and evolutionary search to discover and improve algorithms. Reported applications include faster matrix multiplication, improvements to computing infrastructure and solutions to mathematical problems. By 2026, Google reported further work involving DNA sequencing, disaster prediction, power-grid simulations and scientific computing.

That changes the game slightly. Traditional science usually looks something like: human has idea, human tests idea, human gets annoyed, human tries another idea.

AI can potentially run the cycle at enormous scale: generate thousands of hypotheses, test them computationally, discard the rubbish and send the promising candidates to humans or automated laboratories.

And that is probably the real story.

AI has not “solved everything”. It has started solving pieces of problems that once seemed stubbornly resistant to computation, while opening entirely new ways of searching through possibilities.

The next breakthrough may therefore not be one spectacular answer.

It may be the machine that keeps finding the next question.

Sources list:

Nobel Prize — The Nobel Prize in Chemistry 2024 — https://www.nobelprize.org/prizes/chemistry/2024/summary/

Google DeepMind — Millions of new materials discovered with deep learning — https://deepmind.google/blog/millions-of-new-materials-discovered-with-deep-learning/

Google DeepMind — AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms — https://deepmind.google/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/

Google DeepMind — AI achieves silver-medal standard solving International Mathematical Olympiad problems — https://deepmind.google/blog/ai-solves-imo-problems-at-silver-medal-level/

Google — AlphaEvolve, 1 year later: Impact on science, technology — https://blog.google/innovation-and-ai/infrastructure-and-cloud/google-cloud/alphaevolve-updates/

MIT News — Artificial intelligence yields new antibiotic — https://news.mit.edu/2020/artificial-intelligence-identifies-new-antibiotic-0220

MIT Jameel Clinic — Antibiotic identified by AI — https://jclinic.mit.edu/antibiotic-identified-by-ai/

MIT Collins Lab — Antibiotics-AI Project — https://www.collinslab.mit.edu/antibiotics-ai-project

Thisclaimer — https://thisclaimer.com

What Has AI Actually Solved? The Biggest Problems Where Machines Found Real Answers

Dear Cherubs, AI has been promised as everything from an office assistant to the machine that will apparently fix civilisation before lunch. Strip away the hype, though, and something genuinely remarkable has happened: AI has already helped solve, or substantially crack, several problems that humans had struggled with for decades.

The important bit is “helped”. AI has not cured cancer, solved climate change or discovered the meaning of life while we were making coffee. But in science, mathematics, medicine and engineering, it has produced results that have survived contact with reality.

PROBLEMS THAT GOT REAL ANSWERS

Perhaps the most famous example is protein structure. For decades, predicting how a protein folds into its three-dimensional shape was a major biological challenge. DeepMind’s AlphaFold transformed the field by predicting protein structures at unprecedented scale, contributing to work that earned Demis Hassabis and John Jumper the 2024 Nobel Prize in Chemistry.

Then there are antibiotics. MIT researchers used machine learning to identify halicin, a previously known compound with powerful antibacterial activity, including against several drug-resistant bacteria. Later work identified abaucin, a new antibiotic candidate targeting Acinetobacter baumannii, with activity demonstrated in laboratory experiments and a mouse infection model. MIT’s Antibiotics-AI project now reports nine novel antibiotics discovered or designed with AI and more than 70 billion molecules screened computationally.

Materials science has produced another eye-watering number. Google DeepMind’s GNoME system predicted around 2.2 million crystal structures, including 380,000 predicted to be particularly stable. Researchers independently produced 736 of those predicted structures experimentally, turning some computer-generated possibilities into actual materials.

Mathematics has also taken a hit from the robot army. In 2024, DeepMind reported that AlphaProof and AlphaGeometry solved four of six International Mathematical Olympiad problems, reaching silver-medal-level performance. AlphaProof also produced formal proofs for its solutions, rather than merely throwing suspiciously confident numbers at the wall.

FROM ANSWERS TO DISCOVERY

The more interesting development may be that AI is increasingly being used not simply to answer questions, but to generate new solutions.

AlphaEvolve, introduced by Google DeepMind in 2025, combines large language models with automated evaluation and evolutionary search to discover and improve algorithms. Reported applications include faster matrix multiplication, improvements to computing infrastructure and solutions to mathematical problems. By 2026, Google reported further work involving DNA sequencing, disaster prediction, power-grid simulations and scientific computing.

That changes the game slightly. Traditional science usually looks something like: human has idea, human tests idea, human gets annoyed, human tries another idea.

AI can potentially run the cycle at enormous scale: generate thousands of hypotheses, test them computationally, discard the rubbish and send the promising candidates to humans or automated laboratories.

And that is probably the real story.

AI has not “solved everything”. It has started solving pieces of problems that once seemed stubbornly resistant to computation, while opening entirely new ways of searching through possibilities.

The next breakthrough may therefore not be one spectacular answer.

It may be the machine that keeps finding the next question.

Sources list:

Nobel Prize — The Nobel Prize in Chemistry 2024 — https://www.nobelprize.org/prizes/chemistry/2024/summary/

Google DeepMind — Millions of new materials discovered with deep learning — https://deepmind.google/blog/millions-of-new-materials-discovered-with-deep-learning/

Google DeepMind — AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms — https://deepmind.google/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/

Google DeepMind — AI achieves silver-medal standard solving International Mathematical Olympiad problems — https://deepmind.google/blog/ai-solves-imo-problems-at-silver-medal-level/

Google — AlphaEvolve, 1 year later: Impact on science, technology — https://blog.google/innovation-and-ai/infrastructure-and-cloud/google-cloud/alphaevolve-updates/

MIT News — Artificial intelligence yields new antibiotic — https://news.mit.edu/2020/artificial-intelligence-identifies-new-antibiotic-0220

MIT Jameel Clinic — Antibiotic identified by AI — https://jclinic.mit.edu/antibiotic-identified-by-ai/

MIT Collins Lab — Antibiotics-AI Project — https://www.collinslab.mit.edu/antibiotics-ai-project

Thisclaimer — https://thisclaimer.com

What Has AI Actually Solved? The Biggest Problems Where Machines Found Real Answers

Dear Cherubs, AI has been promised as everything from an office assistant to the machine that will apparently fix civilisation before lunch. Strip away the hype, though, and something genuinely remarkable has happened: AI has already helped solve, or substantially crack, several problems that humans had struggled with for decades.

The important bit is “helped”. AI has not cured cancer, solved climate change or discovered the meaning of life while we were making coffee. But in science, mathematics, medicine and engineering, it has produced results that have survived contact with reality.

PROBLEMS THAT GOT REAL ANSWERS

Perhaps the most famous example is protein structure. For decades, predicting how a protein folds into its three-dimensional shape was a major biological challenge. DeepMind’s AlphaFold transformed the field by predicting protein structures at unprecedented scale, contributing to work that earned Demis Hassabis and John Jumper the 2024 Nobel Prize in Chemistry.

Then there are antibiotics. MIT researchers used machine learning to identify halicin, a previously known compound with powerful antibacterial activity, including against several drug-resistant bacteria. Later work identified abaucin, a new antibiotic candidate targeting Acinetobacter baumannii, with activity demonstrated in laboratory experiments and a mouse infection model. MIT’s Antibiotics-AI project now reports nine novel antibiotics discovered or designed with AI and more than 70 billion molecules screened computationally.

Materials science has produced another eye-watering number. Google DeepMind’s GNoME system predicted around 2.2 million crystal structures, including 380,000 predicted to be particularly stable. Researchers independently produced 736 of those predicted structures experimentally, turning some computer-generated possibilities into actual materials.

Mathematics has also taken a hit from the robot army. In 2024, DeepMind reported that AlphaProof and AlphaGeometry solved four of six International Mathematical Olympiad problems, reaching silver-medal-level performance. AlphaProof also produced formal proofs for its solutions, rather than merely throwing suspiciously confident numbers at the wall.

FROM ANSWERS TO DISCOVERY

The more interesting development may be that AI is increasingly being used not simply to answer questions, but to generate new solutions.

AlphaEvolve, introduced by Google DeepMind in 2025, combines large language models with automated evaluation and evolutionary search to discover and improve algorithms. Reported applications include faster matrix multiplication, improvements to computing infrastructure and solutions to mathematical problems. By 2026, Google reported further work involving DNA sequencing, disaster prediction, power-grid simulations and scientific computing.

That changes the game slightly. Traditional science usually looks something like: human has idea, human tests idea, human gets annoyed, human tries another idea.

AI can potentially run the cycle at enormous scale: generate thousands of hypotheses, test them computationally, discard the rubbish and send the promising candidates to humans or automated laboratories.

And that is probably the real story.

AI has not “solved everything”. It has started solving pieces of problems that once seemed stubbornly resistant to computation, while opening entirely new ways of searching through possibilities.

The next breakthrough may therefore not be one spectacular answer.

It may be the machine that keeps finding the next question.

Sources list:

Nobel Prize — The Nobel Prize in Chemistry 2024 — https://www.nobelprize.org/prizes/chemistry/2024/summary/

Google DeepMind — Millions of new materials discovered with deep learning — https://deepmind.google/blog/millions-of-new-materials-discovered-with-deep-learning/

Google DeepMind — AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms — https://deepmind.google/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/

Google DeepMind — AI achieves silver-medal standard solving International Mathematical Olympiad problems — https://deepmind.google/blog/ai-solves-imo-problems-at-silver-medal-level/

Google — AlphaEvolve, 1 year later: Impact on science, technology — https://blog.google/innovation-and-ai/infrastructure-and-cloud/google-cloud/alphaevolve-updates/

MIT News — Artificial intelligence yields new antibiotic — https://news.mit.edu/2020/artificial-intelligence-identifies-new-antibiotic-0220

MIT Jameel Clinic — Antibiotic identified by AI — https://jclinic.mit.edu/antibiotic-identified-by-ai/

MIT Collins Lab — Antibiotics-AI Project — https://www.collinslab.mit.edu/antibiotics-ai-project

Thisclaimer — https://thisclaimer.com

What Has AI Actually Solved? The Biggest Problems Where Machines Found Real Answers

Dear Cherubs, AI has been promised as everything from an office assistant to the machine that will apparently fix civilisation before lunch. Strip away the hype, though, and something genuinely remarkable has happened: AI has already helped solve, or substantially crack, several problems that humans had struggled with for decades.

The important bit is “helped”. AI has not cured cancer, solved climate change or discovered the meaning of life while we were making coffee. But in science, mathematics, medicine and engineering, it has produced results that have survived contact with reality.

PROBLEMS THAT GOT REAL ANSWERS

Perhaps the most famous example is protein structure. For decades, predicting how a protein folds into its three-dimensional shape was a major biological challenge. DeepMind’s AlphaFold transformed the field by predicting protein structures at unprecedented scale, contributing to work that earned Demis Hassabis and John Jumper the 2024 Nobel Prize in Chemistry.

Then there are antibiotics. MIT researchers used machine learning to identify halicin, a previously known compound with powerful antibacterial activity, including against several drug-resistant bacteria. Later work identified abaucin, a new antibiotic candidate targeting Acinetobacter baumannii, with activity demonstrated in laboratory experiments and a mouse infection model. MIT’s Antibiotics-AI project now reports nine novel antibiotics discovered or designed with AI and more than 70 billion molecules screened computationally.

Materials science has produced another eye-watering number. Google DeepMind’s GNoME system predicted around 2.2 million crystal structures, including 380,000 predicted to be particularly stable. Researchers independently produced 736 of those predicted structures experimentally, turning some computer-generated possibilities into actual materials.

Mathematics has also taken a hit from the robot army. In 2024, DeepMind reported that AlphaProof and AlphaGeometry solved four of six International Mathematical Olympiad problems, reaching silver-medal-level performance. AlphaProof also produced formal proofs for its solutions, rather than merely throwing suspiciously confident numbers at the wall.

FROM ANSWERS TO DISCOVERY

The more interesting development may be that AI is increasingly being used not simply to answer questions, but to generate new solutions.

AlphaEvolve, introduced by Google DeepMind in 2025, combines large language models with automated evaluation and evolutionary search to discover and improve algorithms. Reported applications include faster matrix multiplication, improvements to computing infrastructure and solutions to mathematical problems. By 2026, Google reported further work involving DNA sequencing, disaster prediction, power-grid simulations and scientific computing.

That changes the game slightly. Traditional science usually looks something like: human has idea, human tests idea, human gets annoyed, human tries another idea.

AI can potentially run the cycle at enormous scale: generate thousands of hypotheses, test them computationally, discard the rubbish and send the promising candidates to humans or automated laboratories.

And that is probably the real story.

AI has not “solved everything”. It has started solving pieces of problems that once seemed stubbornly resistant to computation, while opening entirely new ways of searching through possibilities.

The next breakthrough may therefore not be one spectacular answer.

It may be the machine that keeps finding the next question.

Sources list:

Nobel Prize — The Nobel Prize in Chemistry 2024 — https://www.nobelprize.org/prizes/chemistry/2024/summary/

Google DeepMind — Millions of new materials discovered with deep learning — https://deepmind.google/blog/millions-of-new-materials-discovered-with-deep-learning/

Google DeepMind — AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms — https://deepmind.google/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/

Google DeepMind — AI achieves silver-medal standard solving International Mathematical Olympiad problems — https://deepmind.google/blog/ai-solves-imo-problems-at-silver-medal-level/

Google — AlphaEvolve, 1 year later: Impact on science, technology — https://blog.google/innovation-and-ai/infrastructure-and-cloud/google-cloud/alphaevolve-updates/

MIT News — Artificial intelligence yields new antibiotic — https://news.mit.edu/2020/artificial-intelligence-identifies-new-antibiotic-0220

MIT Jameel Clinic — Antibiotic identified by AI — https://jclinic.mit.edu/antibiotic-identified-by-ai/

MIT Collins Lab — Antibiotics-AI Project — https://www.collinslab.mit.edu/antibiotics-ai-project

Thisclaimer — https://thisclaimer.com

Two-Dimensional Material Now Easier to Manufacture

https://fed.brid.gy/r/https://hackaday.com/2026/09/28/two-dimensional-material-now-easier-to-manufacture/

Two-Dimensional Material Now Easier to Manufacture

https://fed.brid.gy/r/https://hackaday.com/2026/09/28/two-dimensional-material-now-easier-to-manufacture/

Two-Dimensional Material Now Easier to Manufacture

https://web.brid.gy/r/https://hackaday.com/2026/09/28/two-dimensional-material-now-easier-to-manufacture/

Two-Dimensional Material Now Easier to Manufacture

https://fed.brid.gy/r/https://hackaday.com/2026/09/28/two-dimensional-material-now-easier-to-manufacture/

Two-Dimensional Material Now Easier to Manufacture

https://web.brid.gy/r/https://hackaday.com/2026/09/28/two-dimensional-material-now-easier-to-manufacture/

One of the authors pointed me to some supplemental materials that helped me to grasp what is going on in this paper. I think I'm getting it now.

#MaterialsScience #BobbinLace

www.aalto.fi/en/news/turn...

Turning bobbin lace into shape...

One of the authors pointed me to more information about this project. It includes some helpful video that let me grasp it better.

#MaterialsScience #BobbinLace

https://www.aalto.fi/en/news/turning-bobbin-lace-into-shape-shifting-textiles

One of the authors pointed me to more information about this project. It includes some helpful video that let me grasp it better.

#MaterialsScience #BobbinLace

https://www.aalto.fi/en/news/turning-bobbin-lace-into-shape-shifting-textiles

One of the authors pointed me to more information about this project. It includes some helpful video that let me grasp it better.

#MaterialsScience #BobbinLace

https://www.aalto.fi/en/news/turning-bobbin-lace-into-shape-shifting-textiles

One of the authors pointed me to more information about this project. It includes some helpful video that let me grasp it better.

#MaterialsScience #BobbinLace

https://www.aalto.fi/en/news/turning-bobbin-lace-into-shape-shifting-textiles

One of the authors pointed me to more information about this project. It includes some helpful video that let me grasp it better.

#MaterialsScience #BobbinLace

https://www.aalto.fi/en/news/turning-bobbin-lace-into-shape-shifting-textiles

Researchers have developed a novel metal-organic framework (MOF) photocatalyst that harnesses sunlight to efficiently split water and produce clean hydrogen gas.
#Chemistry #MaterialsScience #RenewableEnergyScience #sflorg
https://www.sflorg.com/2026/09/chm09282601.html

Researchers have developed a novel metal-organic framework (MOF) photocatalyst that harnesses sunlight to efficiently split water and produce clean hydrogen gas.
#Chemistry #MaterialsScience #RenewableEnergyScience #sflorg
https://www.sflorg.com/2026/09/chm09282601.html

Researchers have developed a novel metal-organic framework (MOF) photocatalyst that harnesses sunlight to efficiently split water and produce clean hydrogen gas.
#Chemistry #MaterialsScience #RenewableEnergyScience #sflorg
https://www.sflorg.com/2026/09/chm09282601.html

Researchers have developed a novel metal-organic framework (MOF) photocatalyst that harnesses sunlight to efficiently split water and produce clean hydrogen gas.
#Chemistry #MaterialsScience #RenewableEnergyScience #sflorg
https://www.sflorg.com/2026/09/chm09282601.html

Researchers have developed a novel metal-organic framework (MOF) photocatalyst that harnesses sunlight to efficiently split water and produce clean hydrogen gas.
#Chemistry #MaterialsScience #RenewableEnergyScience #sflorg
https://www.sflorg.com/2026/09/chm09282601.html

I am honored to be recognized as an Expert on the UAE Research Map. 🇦🇪🔬

As an Associate Research Scientist at New York University Abu Dhabi, my research in Additive Manufacturing, TPMS Metamaterials, and advanced materials characterization aligns with the region's push for sustainable innovation.

You can view my research profile here: https://uaeresearchmap.ae/user/shahadat-hussain

Open to collaborations in materials informatics and advanced manufacturing!

#MaterialsScience #AdditiveManufacturing #OpenScience #Research