Let's analyze and fact-check the claims made in the text: 1.
Original video: https://vm.tiktok.com/ZNdHop7M2/
Fact checked on: July 31, 2025
Fact Check Analysis
Let's analyze and fact-check the claims made in the text:
1. "AI already solved biology's 50-year impossible puzzle."
- Fact-check: This refers to the protein folding problem, which has been a significant challenge in biology for decades. AlphaFold, developed by DeepMind, made substantial progress in predicting protein structures, which was indeed considered a major breakthrough in 2020.
2. "AlphaFold, probably one of the most successful AI stories to come out in the last few years."
- Fact-check: AlphaFold is widely recognized as a significant achievement in AI and computational biology, receiving extensive coverage and praise in scientific communities and media.
3. "Your body has over 20,000 different proteins."
- Fact-check: This is accurate. The human genome encodes for approximately 20,000-25,000 proteins.
4. "In 2020, Google DeepMind dropped AlphaFold."
- Fact-check: Correct. DeepMind announced AlphaFold's success in the Critical Assessment of protein Structure Prediction (CASP) competition in 2020.
5. "They fed this AI every known protein structure and every physics rule."
- Fact-check: AlphaFold was trained on a large dataset of known protein structures and uses principles of physics, but the exact details of the dataset and training are proprietary.
6. "90% accuracy on some of the hardest predictions."
- Fact-check: AlphaFold achieved a median Global Distance Test (GDT) score of 92.4 across all targets in CASP14, indicating high accuracy in predicting protein structures.
7. "Papers that used to take 5 years were suddenly done in just an afternoon."
- Fact-check: This is an exaggeration. While AlphaFold accelerates protein structure prediction, the process of conducting and publishing scientific research involves many steps beyond structure prediction.
8. "DeepMind gave it all away for free."
- Fact-check: DeepMind released AlphaFold's code and a database of predicted protein structures to the public, which was a significant contribution to the scientific community.
9. "New malaria vaccines, crops that resist climate change, cancer treatments, plastic eating enzymes, all because AI cracked nature's oldest secret."
- Fact-check: While AlphaFold's predictions aid research in these areas, the development of vaccines, crops, and treatments involves complex processes beyond protein structure prediction alone. AlphaFold contributes to these fields but is not solely responsible for these advancements.
Overall, the text accurately describes the significance of AlphaFold but includes some exaggerations regarding the speed and scope of its impact.
1. "AI already solved biology's 50-year impossible puzzle."
- Fact-check: This refers to the protein folding problem, which has been a significant challenge in biology for decades. AlphaFold, developed by DeepMind, made substantial progress in predicting protein structures, which was indeed considered a major breakthrough in 2020.
2. "AlphaFold, probably one of the most successful AI stories to come out in the last few years."
- Fact-check: AlphaFold is widely recognized as a significant achievement in AI and computational biology, receiving extensive coverage and praise in scientific communities and media.
3. "Your body has over 20,000 different proteins."
- Fact-check: This is accurate. The human genome encodes for approximately 20,000-25,000 proteins.
4. "In 2020, Google DeepMind dropped AlphaFold."
- Fact-check: Correct. DeepMind announced AlphaFold's success in the Critical Assessment of protein Structure Prediction (CASP) competition in 2020.
5. "They fed this AI every known protein structure and every physics rule."
- Fact-check: AlphaFold was trained on a large dataset of known protein structures and uses principles of physics, but the exact details of the dataset and training are proprietary.
6. "90% accuracy on some of the hardest predictions."
- Fact-check: AlphaFold achieved a median Global Distance Test (GDT) score of 92.4 across all targets in CASP14, indicating high accuracy in predicting protein structures.
7. "Papers that used to take 5 years were suddenly done in just an afternoon."
- Fact-check: This is an exaggeration. While AlphaFold accelerates protein structure prediction, the process of conducting and publishing scientific research involves many steps beyond structure prediction.
8. "DeepMind gave it all away for free."
- Fact-check: DeepMind released AlphaFold's code and a database of predicted protein structures to the public, which was a significant contribution to the scientific community.
9. "New malaria vaccines, crops that resist climate change, cancer treatments, plastic eating enzymes, all because AI cracked nature's oldest secret."
- Fact-check: While AlphaFold's predictions aid research in these areas, the development of vaccines, crops, and treatments involves complex processes beyond protein structure prediction alone. AlphaFold contributes to these fields but is not solely responsible for these advancements.
Overall, the text accurately describes the significance of AlphaFold but includes some exaggerations regarding the speed and scope of its impact.
Video transcript
Most people have no idea that AI already solved biology's 50-year impossible puzzle. This is AlphaFold, probably one of the most successful AI stories to come out in the last few years. Let me explain it. So your body has over 20,000 different proteins. Each one folds into a specific shape. And if scientists can't predict that shape, it's really hard for them to build any medicine. And for 50 years, this was biology's biggest unsolved puzzle. It took years and millions of dollars just to figure out one protein structure. Then in 2020, Google DeepMind dropped AlphaFold. They fed this AI every known protein structure and every physics rule. And over time, the AI learned to predict folding patterns like a master chess player seeing 20 moves ahead. This AI solved what took PhD students entire careers in just minutes. I mean, we're talking about 90% accuracy on some of the hardest predictions, something that absolutely baffled the best minds for decades. And of course, when this was dropped, the scientific community went absolutely insane. Papers that used to take 5 years were suddenly done in just an afternoon. Even drug companies that spent decades hunting for new medicines suddenly had a treasure map in their back pocket. Cancer researchers found new targets overnight, all because an AI gave them the molecular keys they needed. And the biggest part of all of this, without a single level of greed, DeepMind gave it all away for free. This database worth of hundreds of billions in research value, just handed to humanity with no strings attached. And we're already seeing the results from AlphaFold. New malaria vaccines, crops that resist climate change, cancer treatments, plastic eating enzymes, all because AI cracked nature's oldest secret. The protein folding problem that defined an entire generation of biologists, solved by artificial intelligence, faster than you can finish this video. And most people scroll past without realizing they just witnessed one of the biggest leaps in AI history. Follow, and I'll keep you posted.