- 6 days ago
- #technology
- #ai
- #robots
- #innovation
- #science
Get ready for a wild ride through the world of tomorrow! In this episode, we dive into mind-blowing tech: Singapore’s living neuron data center, a Superman robot faster than Usain Bolt, AI-powered drones, atom printers for next-gen chips, and a personalized mRNA cancer vaccine breakthrough. Want more awesome tech? Subscribe and tell us in the comments which story blew your mind! #technology #AI #robots #innovation #science
👉 This channel was created in collaboration with https://www.youtube.com/@realidadeimpressionante
🔊🎧Music Info: Provided by the channel "MokkaMusic" and https://inaudio.org
0:00 - Episode Introduction
0:55 - Singapore's Biological Data Center
4:07 - Personalized mRNA Cancer Vaccine Breakthrough
6:34 - Atom Printer and AI-Driven Materials Discovery
9:11 - China's Cliffside Elevator Transformation
11:34 - Humanoid Robot Advances and Competitions
17:22 - AI-Powered Autonomous Drone Technology
👉 This channel was created in collaboration with https://www.youtube.com/@realidadeimpressionante
🔊🎧Music Info: Provided by the channel "MokkaMusic" and https://inaudio.org
0:00 - Episode Introduction
0:55 - Singapore's Biological Data Center
4:07 - Personalized mRNA Cancer Vaccine Breakthrough
6:34 - Atom Printer and AI-Driven Materials Discovery
9:11 - China's Cliffside Elevator Transformation
11:34 - Humanoid Robot Advances and Competitions
17:22 - AI-Powered Autonomous Drone Technology
Category
🤖
TechTranscript
00:00And in today's episode, among other things, you'll see that Singapore gave 16 million
00:05living human neurons to an artificial intelligence to use as the first biological data center
00:11ever created.
00:12In the United States, a so-called drone comes to life with AI and can already operate on
00:18its own.
00:18Atom Printer, the machine that can create the chips for the next technological revolution.
00:23A Chinese humanoid robot called Superman runs faster than Usain Bolt and jumps two meters
00:29high.
00:30Personalized mRNA vaccine for each individual against cancer achieves historic success in
00:36phase three trial.
00:37Let's explore this and more about new technologies.
00:47Hello, impressed viewers.
00:49I'm Realidade Impressionante, and you're watching Realidade Impressionante.
00:55And researchers in Singapore have just introduced a very different kind of data center.
01:00Instead of relying only on traditional processors, those made of silicon, part of their computing
01:05power comes from living human neurons.
01:08These are grown in a lab, okay?
01:10They weren't taken from anyone.
01:11The prototype was installed at the Life Science Institute of the National University of Singapore,
01:15and brings together 20 biological computers called CL1 developed by the Australian company
01:21Cortical Labs.
01:22Together, these systems use approximately 16 million living human neurons.
01:28Each CL1 unit combines neural cells grown from human stem cells with conventional electronic
01:34components.
01:35The neurons grow on an array of electrodes, thus creating an interface capable of sending electrical
01:40signals to all these cells and also recording their responses.
01:43This way, the system can interact directly with the neurons.
01:47They receive information through electrical stimuli, respond to these stimuli, and can modify
01:52their behavior based on what happens during the interaction.
01:55And it's precisely this capability.
01:57The ability to adapt is what makes biological computing so interesting.
02:00The human brain performs extremely complex tasks, while consuming only about 20 watts of energy.
02:07In contrast, the traditional data centers currently used for artificial intelligence can require enormous amounts of electricity.
02:15The idea is not to immediately replace silicone chips with lab-grown brains.
02:21The goal for now is to investigate whether systems made up of living cells can complement
02:26traditional computers in certain tasks, especially those related to learning and adaptation.
02:32The CL1 was developed precisely to allow for this kind of experiment.
02:36Inside each unit, there is an entire system responsible for keeping the neurons alive.
02:41It controls the temperature, nutrient circulation, as well as the filtration of waste and gases needed
02:47for the cells to keep functioning and stay alive.
02:50Therefore, in practice, these computers need to maintain the biological part of their own hardware.
02:55This project brings together Cortical Labs, the Infrastructure Company Day One, and the
03:00Yonglin Loon School of Medicine at the National University of Singapore.
03:03And putting 20 of these units to work together represents a very significant change in scale.
03:09Until now, most of the computing with cultured living neurons was done in individual experiments
03:14inside laboratories.
03:16This new facility places these units in a structure similar to the one already used by conventional
03:22servers, thus forming what its creators described as the first biologically integrated server rack
03:28operating independently.
03:30The researchers intend to use this system to study artificial intelligence inspired by the
03:35human brain, as well as biomedical modeling, the discovery of new drugs, and also treatments
03:41for neurological diseases.
03:42Yeah, it's a fact that they're still a long way from replacing the huge servers currently in use.
03:57And in this case, my friends, a part of the computer isn't just manufactured, right?
04:02It needs to stay alive.
04:04And that's a little strange, isn't it?
04:07An experimental cancer vaccine was individually developed for each patient and reached a very
04:12important milestone in the United States.
04:14Moderna and Merck announced that their personalized mRNA-based therapy achieved the main goals of
04:21a large phase 3 clinical trial involving patients with melanoma, one of the most dangerous forms
04:27of skin cancer.
04:28The treatment, called intismarin autogene, and previously known as mRNA-4157, or V940, doesn't
04:37work exactly like a traditional vaccine, which is intended to prevent a person from developing
04:42a disease.
04:42It is produced specifically for each patient's cancer.
04:46First, a sample of the tumor is genetically analyzed to identify the mutations that are
04:51unique and present in the cancer cells.
04:54Based on this information, a synthetic mRNA molecule is created that can encode up to 34
05:00neoantigens specific to that tumor.
05:03When administered to the patient, the idea is to teach the immune system itself to recognize
05:08these particular characteristics and attack the cells that carry the same mutations.
05:13In the new study, 1134 patients with cutaneous melanoma participated.
05:19All of them had already undergone surgeries to completely remove their tumors.
05:23And the first results from phase 3 were very positive.
05:26According to Moderna and Merck, adding the personalized therapy to Keytruda produced a statistically significant
05:32and clinically relevant improvement in recurrence-free survival.
05:36The study also achieved an important secondary objective related to the amount of time patients
05:42remained without developing metastases.
05:44In other words, it's not a cure yet.
05:47Among patients treated with a combination, the cancer took much longer to return or spread
05:52compared to treatment with Keytruda alone.
05:55The companies have not yet released the full numbers from this phase 3.
05:59Therefore, it is still not possible to know exactly how great the benefit observed in this new study is.
06:04But the earlier results had already drawn much attention.
06:07In phase 2, involving a smaller group of patients, the combination reduced this risk of recurrence or death by 49
06:15after approximately five years of follow-up.
06:17The results, the more detailed results, should be presented at a medical conference
06:21and submitted for scientific publication, while Moderna and Merck intend to discuss the data with regulatory authorities.
06:28We still don't know what the actual effect was, but it's something positive according to what they said.
06:33Yes, artificial intelligence can already predict new materials with promising properties.
06:38And that would be extremely beneficial for society, right?
06:42The problem is that these predictions still need to go through a much more complicated stage.
06:47Turning what exists in a computer with artificial intelligence into a physical material that can actually be manufactured and tested.
06:55The Danish company Adelan 3D wants to reduce exactly this gap with the Nano Fabricator Pro,
07:01a new semiconductor manufacturing platform designed to produce and test materials at the atomic scale.
07:08The company presents the equipment as the first physical platform of its kind, aimed at AI-driven materials discovery.
07:16Currently, even when AI finds a potentially very interesting combination,
07:22researchers still need to produce the material under controlled conditions.
07:26And then, check if it really shows the characteristics predicted by AI.
07:32And these steps usually happen in separate processes.
07:35At the center of the machine is a proprietary technology from Adelan 3D itself, which is called DAUP.
07:42This one allows control of material deposition layer by layer in specific regions,
07:46thus creating extremely small structures, without relying exactly on the same traditional flow
07:52used in conventional semiconductor manufacturing.
07:55However, this still doesn't mean that this machine is now going to replace
08:00the lithography and EUV equipment used in the production of the most advanced chips.
08:04Its goal is a bit different.
08:05To accelerate research into new materials, interfaces, and devices,
08:10quickly turning digital ideas into physical experiments that can be analyzed and tested.
08:15The platform was developed for applications that include semiconductors,
08:19advanced chip packaging, photonics, quantum technologies, and other emerging materials.
08:25Another step planned by Tantum Adelan 3D is to integrate metrology systems directly into the process.
08:31With this, the machine itself will be able, in the future,
08:34to measure what is being produced during its own experiment.
08:37This data could be sent back to the artificial intelligence systems,
08:40which would analyze the results and suggest new combinations or changes in the process.
08:44In practice, this paves the way for increasingly autonomous laboratories,
08:49in which an artificial intelligence proposes the material,
08:52a machine produces the physical structure,
08:54the instruments verify the result,
08:56and this information is sent back again to the artificial intelligence.
09:00And the Nanofabricator Pro is also being prepared for industrial semiconductor settings,
09:04and is manufactured in the United States,
09:06in partnership with Automated Industrial Robotics.
09:11This one is not just tech news,
09:13but also something interesting and human, you could say.
09:17A new elevator built on a huge cliff in southeastern China
09:20is transforming the daily life of residents in a small village isolated among the mountains there.
09:26Standing 288 meters tall,
09:28the new elevator was installed in the Nisru River Grand Canyon in Yunnan province.
09:32It allows people to overcome, in just a few seconds,
09:36an altitude difference that, for many years,
09:39made it extremely difficult for those living in that region to get around.
09:42Among the main beneficiaries are the children of the village.
09:46Before this infrastructure was built,
09:48students and their families had to travel a very steep path through the mountains.
09:52The idea was to reach the lowest part of the valley,
09:55to the area where their school is located.
09:57The journey could take about three hours one way,
10:01meaning three hours to go and three hours to return.
10:04And it involved rocky terrain and very steep sections.
10:08Now the situation is completely different.
10:10The students used a combination of vehicles,
10:12elevators, and cable cars to cross this region,
10:16thus reducing the trip to school to about 30 minutes.
10:19The new elevator, standing 288 meters tall, is not alone there.
10:23It operates alongside a previous structure,
10:26approximately 268 meters tall, which was inaugurated in 2022.
10:31The entire system was built next to the huge rocky walls of the canyon,
10:35and besides serving the residents,
10:36it also ended up becoming an attraction for visitors who come to the region.
10:40And if you're wondering, yes, there are no tolls there.
10:44Local residents can use the infrastructure for free.
10:47The new structure also stands out for its capacity.
10:50The elevator can transport about 1,000 to 1,200 people per hour,
10:55allowing not only the residents of the small communities there to move around,
10:58but also the growing number of tourists who are interested in the canyon's scenery.
11:02The construction is said to have cost approximately $9.5 million when converted.
11:07But perhaps the most interesting impact of all this is not in the size of the structure.
11:11For several generations, geographic isolation has been a part of life
11:15for the families living in this mountainous region in China.
11:19The very location of the community is tied to its history.
11:22The ancestors of these residents are said to have settled deep in the mountains
11:25precisely to protect themselves from conflicts,
11:28thus creating a community that is naturally difficult to access.
11:32That explains it, doesn't it?
11:34And a humanoid robot had a spectacular fall during a training session in Beijing,
11:39just a few days before the start of the second edition of the World Humanoid Robot Games.
11:44The robot was simply there doing a speed test on a track
11:48when it lost control and crashed hard into a cushioned barrier placed at the end of the course.
11:54The impact was strong enough to spark and make parts of the robot break apart as it hit the ground.
11:59It was pretty intense.
12:02The incident happened on August 17th, during the preparations for this competition.
12:06The video quickly started circulating on social media, racking up millions of views.
12:12The cushioned barrier wasn't there by accident, you see.
12:15It had been installed precisely because other robots had already gone past the end of the track.
12:20Yeah, my friends, this episode shows the challenges that engineers still face
12:24when trying to make humanoid robots move at increasingly higher speeds.
12:28Running fast is one thing, but being able to slow down, keep your balance, and stop safely?
12:34That's a whole different story.
12:36The artificial intelligence expert Alan Chen stated that the Chinese robotics industry
12:41might be over-optimizing, so to speak, their robots for speed,
12:45while aspects like stability and braking capability still need more attention.
12:50And it's exactly these limitations that will be put to the test again
12:54at the World Humanoid Robot Games held in Beijing from August 22 to 26.
12:59The scale of the event has grown considerably in just one year, okay?
13:03The 2026 edition will have 2,056 robots, representing 666 teams from 16 countries.
13:11Altogether, there will be 51 events and 1,301 competitions held over five days.
13:18In addition to the traditional races, the robots will also take part in long jump,
13:23weightlifting, tug-of-war, table tennis, and other events.
13:27But the competition wasn't started just to see which machine is fastest or jumps farthest.
13:32Some of the challenges try to reproduce situations that these robots might face in our real world,
13:37including tasks in homes, hotels, industrial facilities, rescue operations, logistics,
13:42libraries, stores, offices, and even vehicle charging.
13:46I just don't know why they're running so much for those kinds of tasks.
13:50I think they're running too much to chase us, right?
13:53Leave it here in the comments, okay?
13:54The rules have also become much stricter.
13:57In the track challenges and other competitive categories,
13:59the robots must operate autonomously, that is, on their own, with AI controlling them.
14:05No humans controlling them remotely,
14:07with the exception, of course, of some specific challenges with obstacles.
14:10For now, that's how it's going to be.
14:12And this turns seemingly simple events, like a race, for example,
14:17into very demanding tests for perception, balance, and motion control systems.
14:21And speed will be precisely one of the most closely watched areas.
14:25In recent months, Chinese companies have been presenting humanoids capable of running faster and faster,
14:31while competitions like these are taking these machines into conditions very different
14:34from the carefully controlled demonstrations that used to take place inside laboratories.
14:39Remember Boston Dynamics?
14:41It's pretty different now, isn't it?
14:44And the Chinese company Unity Robotics has introduced a new humanoid robot
14:48that, according to the company,
14:50is already able to surpass some of the physical limits previously reached by humans,
14:55like Usain Bolt, for example.
14:57Nicknamed Superman,
14:58the new prototype was developed to perform extremely dynamic movements
15:02and managed to reach a top speed of 12.66 meters per second,
15:06which is about 45.6 kilometers per hour.
15:09You can't outrun a machine like that.
15:12For comparison, during his historic 100-meter race in 2009,
15:17Usain Bolt reached a top speed estimated at about 12.42 meters per second.
15:22This means that, at least considering the numbers released by Unity,
15:26the new robot managed to reach a peak speed
15:29that is slightly higher than that of the fastest man in the world.
15:32But running wasn't the only test.
15:34In a demonstration released by the company,
15:36Superman also performed a vertical jump of about 2 meters,
15:40starting from a standstill.
15:41Pretty soon, this thing will be flying.
15:43With a name like that,
15:45Super Superman?
15:46Yeah, and the height is particularly impressive
15:49when we consider that the robot uses legs that are only 85 centimeters long.
15:54And maybe one of the most interesting details
15:56is the time it took to get to this point, you know?
15:59According to Unity,
16:00the prototype was developed in just over three months.
16:03Even so, the company states that it is still in the development process
16:07and that there is plenty of room to improve its performance in the coming months.
16:12That's dangerous!
16:14Superman was not created, at least at this point,
16:17as a domestic robot or a machine intended to work in factories.
16:21It mainly functions as an experimental platform
16:23to study high dynamic humanoid motion.
16:26The higher the speed, the harder it becomes to maintain balance.
16:30As shown in the previous report,
16:32maintaining balance gets harder as speed increases.
16:35During a race, the system needs to continuously calculate the position of the legs,
16:39the contact with the ground,
16:41and the distribution of its weight,
16:43correcting the movement in real time
16:44to prevent the machine from simply losing control.
16:47And this is an area where they've been making progress really fast, you know?
16:51Its previous robots had already demonstrated running,
16:54jumping, kicking, acrobatic movements,
16:57and different ways of regaining balance.
16:58And the new prototype takes this pursuit of physical performance to another level.
17:03However, for now, the numbers presented by Superman were released by Unitree itself
17:08and still need to be evaluated very carefully
17:11when compared directly to human records achieved under controlled sports conditions, right?
17:16But this prototype shows just how far engineers have already managed to push these machines' locomotion.
17:22DefendEye, a drone technology company operating in Poland and also in the United States,
17:27has just received new investment to expand the production of a system that's quite different from conventional drones, okay?
17:33Called the Overwatch drone, it was developed mainly for emergency teams
17:37and can be launched into the air in just a few seconds without the need for a trained human pilot
17:42or even a traditional remote control.
17:45The idea is relatively simple, okay?
17:47In situations like accidents, fires, searches for missing people, or other emergencies,
17:53teams usually need to get to the location, prepare the drone, and rely on someone qualified to pilot it.
17:58DefendEye wants to eliminate a big part of that process.
18:01Their drone is stored inside a tube that's resistant to weather conditions.
18:05When necessary, you just need to activate the system and the device can be launched in less than 10 seconds.
18:10From there, most of the work happens autonomously.
18:13In other words, there's an artificial intelligence embedded in the little guy.
18:17This allows the drone to search for people on the ground, identify their location, and track their movements
18:23while transmitting live images to a command center, which, for now, is operated by humans.
18:29This is particularly useful when the device is monitoring an area from above.
18:33Identifying a person from an aerial image can be difficult, especially at night.
18:37In complex terrains, it's even more difficult, or during an emergency situation, things get really complicated.
18:43That's why this system uses AI to automatically highlight people and images.
18:48The drone also has a camera designed for low light and infrared illumination,
18:52allowing it to detect people even in extremely dark environments.
18:55You can't hide. I'm telling you, no one listens to me.
18:59And it doesn't necessarily depend on GPS to work, okay?
19:03The system was designed for situations where the available infrastructure may be limited,
19:08which is important during natural disasters or operations carried out in mountainous regions.
19:12And there are also configurations where the launch tubes can be permanently installed on fences,
19:18buildings, vehicles, or other structures.
19:21In this case, external sensors can automatically trigger the drone when an event is detected.
19:27The company's proposal is to turn the drone into a tool where artificial intelligence can take over most of the
19:32operation.
19:33And human teams, people, will receive an aerial view of what's happening even before they arrive at the scene.
19:38If you enjoyed this episode, go ahead and leave a like, check that you're subscribed.
19:42And remember that there are two suggestions on your screen.
19:45Just click on one of them, and I'll see you in the next one, alright?
19:48I'll be waiting for you there. Click, and you'll really help me out.
19:51Thanks, bye, see you soon.
19:53I'm waiting, okay? Click there.
Comments