Peter Chen's Profile

I think the goal of self-driving car is not to replace human but to improve the overall efficiency of the society. We used to have 90% of population working as farmer and now that number is less than 1% in U.S. while the society’s productivity and quality of life is much improved. We also witness in the early 20th centuries many home appliances (such as washing machines) saved hours of time for housewives so many of them lost jobs as full-time housewives or cloth washer for other families to become working professionally.

Currently in the Uber ride fare, I believe 60~80% is due to driver salary (depending on the country). Therefore self-driving has the potential to lower the fare to 1/3 or even 1/5 of current level. My argument is that at such price, no one wanna own private car any more therefore Uber will have to buy their own fleet. The operating cost of maintaining such fleet will be much lower than what we currently envision too, for example the cars can automatically park and recharge during the off-peak hour and self-diagnose many issues it detect.

I like your concern in the end that driver may be hesitant to collaborate in the development of driver-less car. However I feel most people do not think too far beyond their direct benefit. Earning some money now vs potentially losing job years in future, I am sure enough people will be interested if the price is right.

Do people still ride house for commuting after automobile is invented? Yes, but for a short period of time (maybe 20 years, from 1890 to 1910) and eventually it becomes an expensive leisure and sport. Same applies to manual vs auto transmission. Most racing cars still have manual but it is becoming a luxury for fewer people.

If self-driving can be intelligent enough to understand human driver perfectly while be more responsive of the surrounding environment, it can coexist with human driver and be more efficient and safer. While I do agree the benefit of self-driving car will be limited if cars with drivers are still on the road, the benefit can be big enough to incentivize people choosing self-driving car. I don’t think we need to ban car with drivers. The saving on insurance, convenience and fuel should gradually make cars with driver obsolete.

Hi Joey,

You did an awesome job bringing up Tesla technology and recent fatality. To be fair, Tesla’s auto-pilot feature is rated as level 2 assisted driving and is never intended to be used in fully-autonomous mode (level 4). It is partially due to Tesla’s mis-advertising to make people believe it can self drive and people rely on it recklessly. Tesla’s Mobileye solution costs around $1,000 vs Google and Uber’s solution costs around $100,000 (due to the expensive Lidar) so it is no way representing what the truly autonomous vehicle’s capabilities. That being said, even with Tesla’s lousy solution, it already achieved a fatality rate lower than current US driver. In US, on average there is one fatality for every 100 million miles and Tesla already accumulated around 200 million auto-pilot miles with one fatality. This number is not statistically significant yet but we can only hope the next gen technology will be much safer.

Regarding to highway safety, it is actually found that driving on highway is much safer and simpler than driving local. Google started road test on highway first and then move to local. It is more predictable environment with less variance. The cognitive load is also simpler on highway than local as objects and intentions are easier to understand. So industry in general does not worry much about highway compared with local environment for self-driving cars.

Actually Tesla is already updating their software regularly to push for new navigation and auto-pilot feature. It is just like iPhone software update and the frequency is much higher than once a year, roughly every month they will drop a new build. Obviously we expect Uber or any other self-driving car company will have this OTA(Over the air) upgrade capability.

Hi Brittany,

Thank you for bringing up the topic of cyber-security. It is indeed one of the biggest concerns for IoT in general. One idea in security is about attack surface, total sum of the vulnerabilities in a given computing device or network that are accessible to a hacker. The attack surface definitely increased over the past decades as we are becoming more mobile connected and more data are stored and accessible from the internet. Although both public and private sectors are putting lots of efforts on cyber security, I feel that the growth of IoT and other connected devices is faster than the development of cyber security.

Short term approach would be increase the investment or regulation around cyber security, making it a mandate for many critical services. Long term I expect industry and research come up with intrinsically more secure mechanism to decrease the vulnerabilities, e.g. block-chain and quantum cryptography.

I understand the moral/ethical questions are big concerns of this technology. Many of these questions do not have clear answers even to a human. Different people may answer those differently and the society tend to accept the moral dilemma people face (for example, the court may find a driver killing pedestrian not liable/guilty if he or she is responding to some uncontrollable force in an emergency and his or her own life is in danger). Therefore I feel some of the discussion is a bit unfair towards the technology as we accept the imperfection of human judgement and morality but demand higher moral standard for technology.

This also reminded me some irkness discussed in class. Certain price/tax strategy usually benefits some while harm others. What matters most to me is that the overall social equity or fairness is improved. In the case of self-driving car, if it can indeed reduce the 90% of fatality by taking dangerous drivers off the road and offering more responsive driving using different sensors, I think the net gain is significant enough even if the technology occasionally made mistakes in some tough moral dilemmas.

I’d love to continue the discussion offline and feel free to reach out to me.

I added the “Future Business Model” section which explains why Uber have to own a fleet of self driving cars since fewer people will own cars in future. Many industry experts believe autonomous vehicles will begin with truck and delivery industry as business has strong interest to save cost and no passenger is involved or giving up control. Once the society becomes more comfortable with many autonomous truck and delivery vehicles on the road, it will become easier for general public to accept the self-driving cars.

I am glad you like it. This is an innovative way of using AI but underneath it is a perfect case to tackle: we have too many parameters and complex nonlinear interactions within parameters so AI is best at finding the answer than human. I can see how this tech can be applied to many other areas too.

Yes, actually there is another approach to cooling. I know one startup is trying to rent family’s basement as distributed server room for free. Basically it works in some cold country where the winter is long and family used to spend significant money on heating every year. By giving out the space of the basement, the server in the basement will run at full speed during winter to generate enough heat for the whole house.

Google did list the PUE for different locations here: https://www.google.com/about/datacenters/efficiency/internal/

You will find the location from tropical areas are generally higher than cold areas. Google did mention they are using free cooling in its DC.

I really like your mentioning of Nest. I think it is always the goal of Nest to have everything connected and interacted coherently in a house. Unfortunately Nest as a business wasn’t doing as well as it hoped. We can chat more about Nest’s recent challenge.

They are actually doing part 1 right now. Before they adopt AI in data center, they already built rules that turn off part of DC during low traffic time. What AI differentiate from human rule is that AI can handle very complicated, inter-relative, non-linear system relationships. Basically for any intuitive rule or heuristics, like turning server off when not used, human are doing perfect job. It is the other jobs AI is best suited for.

Growing land use of DC is an important point to mention! Due to the limit of moore’s law, the speed of shrinking transistor size is slowed down so the density of processor or storage is improving slower. The demand for data grows exponentially while the capacity is growing more like linearly now. This can imply some serious problem down the road if not addressed.

Right now the energy bill is the single biggest cost for data center. I think the cost is even higher than their hardware depreciation cost each year. Even if we move to other ‘free’ energy such as solar, the actual cost of generation electricity will still be a major expense of running data center.

I love your idea of grid level load balancing and global optimization. It is not there yet but it is heading for that direction. We need to have all these facilities connected and share their data all the time.

Consolidate individual, smaller server houses into big data center was actually what happened. This change improved the efficiency so much and kept the growth of total power consumption limited while the capacity increased significantly during last decade.

This also aligns very well with the Cloud computing trend when more and more companies are getting rid of their own server and using Cloud server offered by Amazon or Google remotely. If more work and services are running on Cloud, we can expect the efficiency keep improving. I think for either Amazon or Google, keeping data center energy bill low will continue to be the key competitive advantage so to other company, the cost of using Cloud will be much lower than maintaining their own server.

Current cooling system is either time based or thermostats rule based. The ideal AI system could be aggregating all available info like how many people present in the room, each person’s comfort temperate/what each is wearing today, when you expect people arrive and leave. New building will have most sensor and control needed for this kind of application, it is just software isn’t smart enough to make such call. It is coming really fast though.

In the figure 2, I notice there are some drop of greenhouse gas emission during the past decade, especially the emission of SF6. Does this mean ConEd installed cleaning facility to reduce the emission of certain greenhouse gas or was gradually switching to alternative source that has less impact of greenhouse gas? Was the overall electricity generation of ConEd increased or decreased during the same period? It would be useful to find out ConEd’s timeline towards fulfilling regulatory target by 2030 relatively to its generation growth or decline. As the previous commenter mentioned, reducing greenhosue gas isn’t one company’s duty. If we can significantly reduce the demand of electricity for the next two decades, the combined effect will be much more profound.