TRANSKRIBEERI 🔗
WEBVTT
00:00:05.839 --> 00:00:13.679
Is that we bring together not only folks from industry to talk about what's happening, but also academia to talk about their research.
00:00:14.000 --> 00:00:28.719
And I was talking to someone in the hallway, Sam in the hallway, and we were saying, like, in a very nascent market like this, the blend of academia and industry is really important because it really gives us an insight as to what's happening in this space kind of around the corner.
00:00:28.960 --> 00:00:30.960
Without further ado, Winston.
00:00:33.359 --> 00:00:33.840
Thank you.
00:00:34.320 --> 00:00:35.359
Welcome to Taipei.
00:00:35.600 --> 00:00:45.920
It's really my uh great honor and pleasure to uh share the AI transformation um here in Taipei 2025.
00:00:46.159 --> 00:00:55.280
And I think most exciting that this is my first time joining this community, and I'd like to thank Pete and many community members' uh endorsement invitation.
00:00:55.439 --> 00:00:59.840
And these two days and beyond, I'd like to learn more from you folks.
00:00:59.920 --> 00:01:02.079
I think this is a community effort.
00:01:02.399 --> 00:01:12.640
Today's topics, I even though I'm in academic, I'd like to share uh uh roughly what the technology breakthrough in the past uh 12 or 20 years.
00:01:12.879 --> 00:01:25.439
And I think most important is what the technology we learned in which verticals and likely what kind of commercial opportunity we can bring in in the uh coming years.
00:01:25.680 --> 00:02:10.560
So I give some context, and I'm in the academic and working on some um national key AI initiatives in the past few years, but also Pete just mentioned I also have the automotive role, and I learned how this uh the supply chain and how you make hardware software assistance and even meet the requirement and safety requirements, functional safety requirements in the uh regulation in the industry in different uh regions, and also how energies actually uh bring a huge impact in the past and even the coming year, and how AGAI really also plays a critical role in the um energy uh sectors.
00:02:10.800 --> 00:02:20.800
So so that I would uh get my academia and industry experience and share what was the opportunity ahead and especially for HAI.
00:02:20.960 --> 00:02:28.879
And actually, the opportunity is not does not come from a single pillar or single uh vertical that would bring the impact.
00:02:29.759 --> 00:02:48.879
From my experience in the past few years, I think that would be a critical but uh very intertwined uh pillars ahead for H AI as AI energy and autonomy as as well.
00:02:49.199 --> 00:02:49.759
Why?
00:02:49.919 --> 00:03:01.120
I think uh the AI, everybody understands that it's it's quite uh important to bring the intelligence to the device level, but the AI requires tons of energies.
00:03:01.360 --> 00:03:12.560
So energy defines the boundary, defined the limitations of AI capabilities, more tokens, more energy, you have more tokens, you got
Load full transcript
to bring more intelligence.
00:03:12.800 --> 00:03:34.080
But the other hand is AI actually have energy a lot from our experience that the AI on the age side and cross side can give better forecast, predictive maintenance for the grid, for the power generations, even bring the energy uh efficiency in the home or in the factory or in the uh data centers.
00:03:34.240 --> 00:03:37.039
So AI have a lot for energies.
00:03:37.120 --> 00:04:01.280
And also autonomy that no matter its agent, physical agent doing the control or autonomous driving and all the coming opportunity for the robot also require the energy to drive the motor, drive the application, also require the AI intelligence that can have autonomy really operate uh well smoothly in such dynamic and challenge change uh environment environment.
00:04:01.439 --> 00:04:10.319
So I would say that AI, energy, autonomy actually intertwine together, as you you will see that in the uh coming tech.
00:04:10.879 --> 00:04:18.160
It's really hard to predict the future because I always have uh lots of questions from the government, from the industry partner.
00:04:18.720 --> 00:04:20.319
Could you say what should we do?
00:04:20.639 --> 00:04:24.240
What will be the focus, what will happen in the technology sector?
00:04:24.399 --> 00:04:27.759
So that's why I found that the best answers go to the history.
00:04:28.160 --> 00:04:34.000
Here is the curve showing the GDPs of the world in the past uh 200 years.
00:04:34.879 --> 00:04:41.680
This is log scale, even though it's all they linearly growth, but actually it's actually exponential growth here.
00:04:42.319 --> 00:04:56.000
And it's you see that, yeah, this can need growth, this is good for the society, but uh GDP also uh can reflect the economic conditions and also the wealth and also the productivity for the nation or for the world.
00:04:56.240 --> 00:04:58.879
But is really GDP continued to growth?
00:04:58.959 --> 00:05:02.959
And this is one of the questions many people ask, even productivity, you can need to grow.
00:05:03.680 --> 00:05:11.279
But we saw that actually the driving the for the growth for the GDP aligned with the technology uh revolution.
00:05:11.360 --> 00:05:36.079
So you starting for the first industry revolution, starting from uh 1770, is the introduction of water powers that drive the machines that have more efficiently uh threading the uh the cut threading and spinning the cast threads, and also following with in the 80s, in 1800s, there for the introduction of electricity, heavy machine, oil, etc.
00:05:36.399 --> 00:05:46.800
So you can see that the industry revolutions, they bring the energy first, water, electricity, oil, and then drive the automation.
00:05:47.199 --> 00:05:58.000
Even though it's not fully autonomy yet, but I think this intention that they try to uh drive the the machine uh automation that reduces the human power, improve the productivity.
00:05:58.079 --> 00:06:01.519
That's the same as what we want to expect for the autonomy.
00:06:01.680 --> 00:06:12.079
So you can see that in the past, the energy autonomy determined the productivity, determined the competition, determined the GDPs for the country, for the enterprise.
00:06:12.319 --> 00:06:29.199
And then uh following the the last uh innovation or technology uh uh improvement in 1971, the ID technology we generate more data, more machines, more competition, it gets into a digital war, and then following the last uh revolution in AI revolution.
00:06:29.360 --> 00:06:37.040
So we can see that energy automation and then following the AI will be the next right for the coming uh war.
00:06:37.360 --> 00:06:40.079
So we need to pay more attention on that.
00:06:40.240 --> 00:06:45.120
And then look at first why AI would be a good fit for HAI.
00:06:45.199 --> 00:06:52.319
But I want to emphasize that the automation, autonomy, energy, AI intertwine all together.
00:06:52.480 --> 00:06:55.920
You should look then in the holistic way.
00:06:56.720 --> 00:07:04.319
So these curves just summarize the AI technology uh development in the past uh 12 uh years.
00:07:04.560 --> 00:07:08.560
We have been working on these AI technologies since 2002.
00:07:08.639 --> 00:07:09.920
It's 23 years.
00:07:10.000 --> 00:07:15.759
So we can easily sense the dramatic change in the past 23 years.
00:07:16.000 --> 00:07:23.920
So the curve showing that is the growth of the number of parameters in different uh major parameters.
00:07:24.079 --> 00:07:35.279
So you can see the curve starting from 2013 and continued growth, the number size of parameters starting from a median scale to billion scale, even trillion scale right now.
00:07:35.360 --> 00:07:41.439
But we can say that the more uh parameters, the power the more powerful the model.
00:07:42.000 --> 00:07:46.319
But generally, generally we can categorize that into four stages.
00:07:46.480 --> 00:08:14.720
The first stage is regarding the procession, it means that at the time the neon world, uh convolutional neuronal world, redominate the uh the the academic or industry, so all the convolution technology you need every developed uh at that stage, and then introduction of transformer, then the natural language processing community understands how to use the technology to uh improve the uh Q and A translation, etc.
00:08:15.199 --> 00:08:39.679
And later people find that transformer is easy to scale to billion scale pair merger, then the capability emerge and and the introduction of uh GPT-3 in uh 2020 and also the end of 2022 for the uh GPT is really uh bring quite impact to the uh industry and also the academia as well.
00:08:39.759 --> 00:08:40.799
Why did the impact?
00:08:41.039 --> 00:08:46.159
Because it carries very, very important uh uh property co-generation.
00:08:46.320 --> 00:08:51.200
It means that one was the model trained in certain conditions, they can generate well too many occasions.
00:08:51.360 --> 00:09:01.360
This kind of generation
capability, and also we found that the model also carries some uh uh task or capability that within the single model.
00:09:01.440 --> 00:09:03.679
This has never happened uh uh before.
00:09:03.840 --> 00:09:12.000
So, generation, my tasking is bringing lots of imagination for the industry, seeing that this model can do lots of things.
00:09:12.399 --> 00:09:16.080
That's why we saw the huge growth in this AI ecosystem.
00:09:16.240 --> 00:09:21.519
Normally it's the OE model, even how system age as well, we impact a lot.
00:09:21.759 --> 00:09:31.919
But we said that the the most thing we should uh care about is the four-stage reasoning uh capability that just emerged from last year.
00:09:32.320 --> 00:09:46.320
When you use more token, more time to do in the inference, you can try to deal with very challenged problems, like system two reasoning and quitting thinking, science, math, physical problem, uh, and AI for science as well.
00:09:46.559 --> 00:09:56.320
You think about that, it's not only the chipbot, but also this kind of quitting thinking, training, can lending from the crowd from the crowd to the edge.
00:09:56.480 --> 00:09:59.279
What other application you can do you can bring in?
00:09:59.360 --> 00:10:10.399
You can think that you have some assistant along with you at home, at very private sections, and even in lots of a critical area, you have a thinker go with you.
00:10:10.559 --> 00:10:14.159
It will enable lots of exciting applications.
00:10:14.399 --> 00:10:18.240
But one question people have is this is a building scale.
00:10:18.399 --> 00:10:22.399
Is even trillion scale model, we're resource for that.
00:10:22.480 --> 00:10:26.240
Is it really possible that we can bring that into the SI?
00:10:26.799 --> 00:10:30.960
One year before I say, I don't know, but when I would say, yes.
00:10:31.200 --> 00:10:40.559
I think you already experienced, I call the deep sick incident, uh, I think in early January 2025, uh this year.
00:10:40.799 --> 00:10:49.279
So this is uh uh a company that brings lots of uh uh noise or brings lots of tensions in early 25.
00:10:49.360 --> 00:10:56.879
But I think this is not a technology breakthrough, but uh we said that from the academic view, it's a great engineer implementation.
00:10:57.039 --> 00:11:03.840
From their proposal, they bring the uh training and inference efficiency into the large energy model.
00:11:04.080 --> 00:11:09.200
And it also intensifies the competition in the large learning model company.
00:11:09.279 --> 00:11:15.919
You see more capacks and efficiency, but also reduces the price for the uh tokens.
00:11:16.159 --> 00:11:23.120
But actually, I would say that the AGI community actually benefits a lot from this uh computation.
00:11:23.279 --> 00:11:44.799
And it's through this competition that you the community, the large learning model community actually accelerates the development of some of small but very powerful models such as large language model and also even
the VLM model that is multimodal combined with language, with vision, with speech, speech as as well.
00:11:44.960 --> 00:11:52.080
And also, the model also embeds some, I said, very important quick thinking, planning, reasoning capabilities.
00:11:52.399 --> 00:12:09.279
Most important, this model, we see some small model, a billion and two dozen billion, even a hundred billion model, and actually through many, many experiments show that this billion, hundred billion scale of model is a good fit for the edge device.
00:12:09.440 --> 00:12:13.039
Even three billion model is a good edge device as well as well.
00:12:13.200 --> 00:12:18.000
We don't have time to go through that, but actually lots of powerful model we can work on that.
00:12:18.159 --> 00:12:26.639
But it also starts with uh trigger the decentralization, uh they migrate the technology from the car to edge.
00:12:26.799 --> 00:12:30.639
Actually, this kind of decentralization happened before.
00:12:30.799 --> 00:12:37.519
Uh internet age, that when the packet gets cheaper, it drives the decentralization into the the age.
00:12:37.600 --> 00:12:39.919
It's also right now, token got cheaper.
00:12:40.080 --> 00:12:42.480
I think it's also an opportunity for the age.
00:12:42.639 --> 00:12:53.759
As Pete just mentioned that uh some the application require real-time uh privacy, even low latency, they require uh this kind of technology, and so high growth super role.
00:12:54.000 --> 00:13:06.960
But I see that uh uh from the market public report, we saw that uh, at least from the SOC, uh their uh quoting report, we saw that the smartphone, PC, PC benefit a lot from the HGAI.
00:13:07.039 --> 00:13:13.200
I would say this is just a starting point, it's just a starting point from that, even some ARGS, iPhone, etc.
00:13:13.679 --> 00:13:17.200
And this potentially we see the home appliance.
00:13:17.440 --> 00:13:27.200
Uh, I think uh Peter just mentioned the air conditioned cases as is also in this category, toys, education, office, knowledge management would benefit a lot.
00:13:27.360 --> 00:13:30.000
And right now, lots of people pay attention really.
00:13:30.159 --> 00:13:40.080
Is this possible that this kind of AI can migrate from the digital world into real world and really doing some actions that fall into robot and autonomy?
00:13:40.399 --> 00:13:45.120
And this is with the other pillar we mentioned is autonomy and pillars.
00:13:45.360 --> 00:13:54.240
In this pillar, most people care about the robot taxi and like require a lot of sensors, a lot of computing, and also huge capex investment.
00:13:54.559 --> 00:14:14.559
But honestly, from my experience in the supply chain, in the automobile industry, and also workers with those uh OEMs and global OEMs, we know that the also many reports showing that the volume of Robot Taxi will be few, even in the coming 10 to 2035.
00:14:14.639 --> 00:14:15.840
It's only a small volume.
00:14:16.159 --> 00:14:24.720
Every year there's around
80 million new vehicle manufacturers in the year, but the majority for the autonomy will be L2 Plus.
00:14:25.200 --> 00:14:28.159
It's I think many people already experienced that.
00:14:28.320 --> 00:14:35.679
It's it's partial autonomy means that sometimes you can hand uh you can hands off, but you still have some attention for that.
00:14:35.759 --> 00:14:45.519
But still, this L2 autonomy, plus autonomy already bring lots of convenience for many users along with the uh software define vehicle.
00:14:45.679 --> 00:14:51.759
We need to bring better experience onto the diesel carpet and also some safety requirements.
00:14:51.840 --> 00:14:57.600
So we can say that LO1, LO2, even L2 Plus will occupy huge potentials.
00:14:57.759 --> 00:15:09.519
You so you see that it's kind of transition from the distributed uh issues into centralized SPC, it requires more connection, more memory, more uh uh requirement on that.
00:15:09.919 --> 00:15:23.440
And the automotive experience actually can help a lot for the robot industry as well, because it's timing that how we have this bot can go into the the real world.
00:15:23.600 --> 00:15:27.600
It's not only uh in the digital world for the reasoning, for the perception only.
00:15:27.679 --> 00:15:36.639
It needs to turn to the action, you need to do grasping, it needs to the planning, you need to do navigation, you need to uh um uh interact with the real world.
00:15:36.720 --> 00:15:39.759
But the real world is so dynamic, it's so challenging.
00:15:39.919 --> 00:15:55.279
And there's really quite strong needs, even I I go I involve some national program that because of the aging of people and shortage of labors, the service healthcare logists have strong needs for this kind of robot.
00:15:55.840 --> 00:16:11.120
And the same thing is that if we really can really scale out the deployment for the robot, we need the certain type of generation capability, multitasking capability, as we see from the large language model in the digital world.
00:16:11.360 --> 00:16:13.840
Can we really reach that in the coming years?
00:16:14.000 --> 00:16:16.080
I think they are really quite attempt.
00:16:16.240 --> 00:16:21.759
They try to bring the large language models, VLM model, and to help the robot world.
00:16:21.919 --> 00:16:28.720
And I would say that's the key to unlock these critical capabilities.
00:16:29.039 --> 00:16:32.000
I just showed some few work from my lab.
00:16:32.159 --> 00:16:48.399
So we are working on that uh in the past 10 years and try to uh bring the uh have a smart brand for robots, so we saw that the Last Rail model can help even have prone to uh guide the robots to do some specific tasks.
00:16:48.559 --> 00:17:09.920
And also the VLO model can help not only build a map for the robots, it's helped to navigate the world, even doing the search, but also the current technology can help uh real time build up very, very dense map and keep in mind that these things all can be done, have to be done in the edge side.
00:17:10.000 --> 00:17:12.240
And because this we need to do that in real time.
00:17:12.400 --> 00:17:22.640
And also the large learning models, LLM can also help bring the common sense to the robot, train the robot, and they have some common sense as they navigate in the uh real world.
00:17:22.720 --> 00:17:25.440
And also data is always a big issue.
00:17:25.599 --> 00:17:29.359
How you much data you need to train the robot.
00:17:29.519 --> 00:17:36.559
I think the LM model benefits from using the tons of free data, conducting unsupervised uh trainings.
00:17:36.799 --> 00:17:42.079
Is it possible that in robot world we can also have this kind of unsupervised learning?
00:17:42.240 --> 00:17:48.559
Just take the uh video data and put it in the training space and robot can learn from that.
00:17:48.799 --> 00:17:51.839
Um so I show you uh quite quick demos.
00:17:52.079 --> 00:17:54.079
This is we did it five years before.
00:17:54.319 --> 00:18:00.160
So you can see we use prompt to drive the machine to finish the task we tend to have.
00:18:00.559 --> 00:18:08.160
So they can understand the spatial relationship, um, they can uh grasp any kind of object, uh, even though they don't recognize the object.
00:18:08.400 --> 00:18:10.720
But they already have this kind of capability.
00:18:12.480 --> 00:18:15.599
So this is technology have been, we did it five years before.
00:18:15.680 --> 00:18:19.759
You can imagine that right now, the five years, the large unit model evolve a lot.
00:18:19.839 --> 00:18:21.200
They will bring more impact.
00:18:21.440 --> 00:18:34.799
The final demo I will show you that is it possible to run through this regular video to the YouTube video and try to find the locomotion navigation and try to teach the robot for that.
00:18:34.960 --> 00:18:43.119
And actually, this through two or three videos, the robot can easily learn how to navigate in the real environment, even the robot in different forms.
00:18:43.200 --> 00:18:46.559
So this is called cross embodiment learning.
00:18:46.640 --> 00:18:54.240
So I will show you, yes, that's really quite invest improvement in this technology, help the robot to learn the real world.
00:18:54.400 --> 00:19:15.759
But people will ask questions is really we can bring the large language model, fraudulent model into the robot, and then they can then mitigate in the real world, and also have this kind of generation multitasking capability as the ChatGPT, if that's really great, they will be imposed lots of opportunity for that.
00:19:15.920 --> 00:19:20.160
I would say yes, the community already working a lot for that.
00:19:20.319 --> 00:19:35.839
I think this since the end of uh of 2020, I think also motivated by the ChatGPT that uh started introducing the RT1, RT2, RTX, and also the open community, uh open source community introduced the open VOA.
00:19:36.079 --> 00:19:43.279
Recently, the uh staff called uh Physics AI, they introduced Pi Zero and Pi O5.
00:19:43.680 --> 00:19:46.799
I would say they're really uh uh quite inspiring.
00:19:46.960 --> 00:19:50.240
You see that, you can try to see the Pi O5 demo.
00:19:50.319 --> 00:19:59.359
They try to train the model in the virtual world and bring the the robot into different homes they rented from Airbnb.
00:19:59.920 --> 00:20:05.519
It can follow the comment to clean the kitchen and clean the bedroom, etc.
00:20:05.839 --> 00:20:15.680
So we saw is apparently some generation uh multitasking can be already go into these kind of solutions.
00:20:15.920 --> 00:20:34.319
Even though it's, I think this is still a little distance from the real application, but I think current technology can already help a lot in the closed environment, in the hospital, in the industry, in many side, really bring some really flexibility, intelligence, and efficiency to the robot already.
00:20:34.559 --> 00:20:35.680
We can use that already.
00:20:35.839 --> 00:20:43.839
But the other question is but is it's really a training scale model or even hundreds of something billion a model for that?
00:20:44.000 --> 00:20:49.759
No, this model you can get from open source is three to fifty uh five billion.
00:20:49.839 --> 00:20:52.799
Even the model I mentioned, Pi 05, is three billion only.
00:20:52.880 --> 00:20:59.200
So easily you can take that to uh implement in the HSI or radio.
00:21:00.079 --> 00:21:07.359
But very different from the uh digital world, the ChatGPT sticker they have, uh they will make mistakes.
00:21:07.440 --> 00:21:08.960
They will do hallucinations.
00:21:09.599 --> 00:21:14.400
But the hallucinations in the real world would cause lots and lots of risk.
00:21:14.480 --> 00:21:23.759
You can imagine that if the autonomous driving robot taxi or the robot hallucinate in the real world, what might happen?
00:21:24.000 --> 00:21:30.559
But so that's why in autonomic industry, automatic industry, they have very strong regulation.
00:21:30.640 --> 00:21:33.279
They have a lot of standards, compliance.
00:21:33.440 --> 00:21:43.920
For example, how you make sure the functional safety or the quality in the long term, and also safety in lots of scenarios, in design, and use of cybersecurity and for that.
00:21:44.079 --> 00:21:55.519
And actually, those experiences going through in the automatic industry, standards in the automotive industry can also help a lot for the robot development as well.
00:21:55.839 --> 00:22:03.920
I would say that this is quite early days, it's it's really coming for this uh foundation model into the robot worlds.
00:22:04.079 --> 00:22:11.440
But I would say that the robot will be in many, many kinds of forms that it will be out of our imaginations.
00:22:11.680 --> 00:22:16.799
The next pillar I would talk about is the energies parts.
00:22:17.200 --> 00:22:18.799
Why energies?
00:22:19.039 --> 00:22:47.599
That we have uh quite uh uh surging energy investment because the AI requires energy and the the energy uh sectors uh uh is quite uh uh developing with introduction of renewable energy, the grid needs to be uh uh upgraded and the generation from centralization into uh uh decentralization and also the sustainability needs for IE100 or 24-7 CFE.
00:22:48.079 --> 00:23:02.640
But I wanted to emphasize that the HGAI actually played a very, very critical role to bring the efficiency and also the resiliency for the energy uh network, even energy at home at a uh factory.
00:23:02.880 --> 00:23:13.200
This is the best story that we we all know that in April it happened in Spain and Portugal because of some issue that uh made the uh power outage.
00:23:13.279 --> 00:23:14.480
It caused lots of problems.
00:23:14.640 --> 00:23:27.440
But actually, through many, many analyses, they found that if you bring additional uh battery and also some smart control, intelligent control here, actually we can avoid even mitigate these uh outages.
00:23:27.759 --> 00:23:42.400
That's why uh we also uh we have uh this uh company really right now operate uh management uh 5.5 gigawatt over the world, and we daily we process 1.3 billion data points in the age, also in the cloud.
00:23:42.480 --> 00:23:49.599
We operate uh more than uh, I think it's it's 13 countries uh right now and manage more than 10,000 uh data centers.
00:23:49.759 --> 00:23:55.680
Uh one of that, like this, this is the largest data center, uh the solar panel power plant in Taiwan.
00:23:55.759 --> 00:23:58.720
This is one of the sites we management uh uh right now.
00:23:58.880 --> 00:24:00.720
You can imagine lots of age activities.
00:24:00.799 --> 00:24:04.319
Some it's not only monitoring, sometimes we need to control as well.
00:24:04.480 --> 00:24:07.519
But this control also incurs lots of risk.
00:24:07.839 --> 00:24:34.640
This is a final example I want to show you that that should actually, one of the key uh critical applications either we can use in the in the data centers, use in the factory or use it at home is the so-called battery storages that can try to bring the resilience, even the cost uh reduction uh in the in the in at home or in the factory or in in the uh utility scale data center.
00:24:34.720 --> 00:24:43.440
Then require to integrate the energy system storage with the grid, with the power generation, uh solar, uh wind, or EV charging, battery storage.
00:24:43.599 --> 00:24:48.880
So we need require dynamic inter control, try to optimize, release, optimize the cost.
00:24:49.279 --> 00:24:59.200
And we found that we just pilot a project that this kind of technology requires reinforcement, the one that FIGO use in the HSI.
00:24:59.279 --> 00:25:11.359
So we already deployed that in some of our customers in Japan, that we try to use this AI technology, try to bring the efficiency and resilience for the energy sector.
00:25:11.519 --> 00:25:12.559
This is my last slide.
00:25:12.799 --> 00:25:25.920
So what we I want to highlight that it's just the start of HGAI community, and we found the critical pillar we can uh contribute the most for AGI community is energy, AI and autonomy.
00:25:26.160 --> 00:25:29.279
We need to think that all consider all together.
00:25:29.440 --> 00:25:31.359
You can see there are lots of opportunity.
00:25:31.519 --> 00:25:35.440
We don't have time to cover that, you see lots of opportunity within this one.
00:25:35.839 --> 00:25:42.720
And right now, the technology is already moving from the centralizations, official AI to the distribution.
00:25:42.960 --> 00:25:54.720
And and the I assure you that the current model is already powerful enough and it's is it's engineer-friendly for HAI already.
00:25:55.039 --> 00:26:01.359
But for a company that's designed a product, yeah, we don't know if it really could fit post-customers.
00:26:01.519 --> 00:26:07.839
So my experience is also showing that we need to design lots of POCs and try and error with the system company.
00:26:08.160 --> 00:26:16.079
And but we need to keep in mind that this, when the HDI go into real world, conduct lots of control, lots of action, it's not only a digital world.
00:26:16.160 --> 00:26:18.880
So we need to avoid the hallucination in the real world.
00:26:19.039 --> 00:26:25.039
So the action might turn to some mistakes and risk, and how to mitigate the risk is a huge concern.
00:26:25.200 --> 00:26:32.240
Safety, cybersecurity, and also the other consideration need to be required for SOC or citizen company as well.
00:26:32.319 --> 00:26:39.200
So that's why we have this HGI foundation and we can work together for this opportunity and changes.
00:26:39.359 --> 00:26:40.160
Thank you.