The goal of AutoML is to compress the time/money from idea/problem def to running models, it is not to replace the data scientist.
825 karma · joined August 11, 2017
The goal of AutoML is to compress the time/money from idea/problem def to running models, it is not to replace the data scientist.
I do agree, that any auto ml tool, cannot know what is the correct data sources for a specific use case, nor does it know the business case value of a specific model.
Auto ML should not replace humans, but work with them. I.e. it is a tool.
This tool can do the following:
1. Auto visualization.
2 Auto feature engineering.
3. Auto backtesting.
4. Auto training (with cost optimization).
5. Hyperparameter optimization.
6. Auto model packaging.
7. Autoload testing and auto security testing.
7. Auto deploy / scale / SLA.
8. Auto monitor.
The goal of AutoML is to save time and reduce risk. I am not sure why you want to do all of the above manually.
Which is, vendor locking did not show itself, because the economy was good.
It basically a fixed cost imposed by the cloud providers.
I.e. when the economy contract, and yet AWS/Azure profits are growing, this is the pain of vendor locking.
If yes, is there a way to certify an operator on Vmware 7?
Each language is a much more than its syntax. It is a complete ecosystem with network effects.
Rust can overcome go, if it would provide a lot of open source libraries, as well as growing number of programmers.
To do that, it would need the backing of big tech (e.g. MS), that have a big budget to evangelize it.
In any rate, the choice of language today is completely meaningless, since you can wrap anything in a container, and put a GRPC interface on top of it.
So the choice of lang should be based on the libraries for the task at hand and not on the lang itself.
I actually expect Kubernetes to start offering the AWS services as CRD/operators, and not the other way around.
cloud is dead.
Probably currently see some loss of adaption due to this delay.
I just wonder why not CNCF from the get go.
% of investments exited 484 71 58 77 89 22
I guess that on avg, 70% of startup exit, vs the common belief of failure rate of 90%.
I.e the failure rate is 30%.
I.e. contact tracing is like a BFS search on a graph.
So, even if you trace the closest contacts, by the time a person takes the test and notify the network, those primary contacts (edge 0) might have infected their own contacts (edge 1) until basically you would need to isolate the whole population.
1) https://www.amazon.com/Object-Oriented-Modeling-Design-Rumba...
2) https://www.amazon.com/OMT-Insights-Perspective-Object-Orien...
1) https://www.amazon.com/Executable-UML-Foundation-Model-Drive...
And the work of daniel Jackson from MIT :
http://people.csail.mit.edu/dnj/
Look at this course, at the videos at the bottom:
https://stellar.mit.edu/S/course/6/fa18/6.170/materials.html
Jackson is also behind Alloy:
https://mitpress.mit.edu/books/software-abstractions-revised...
The question is not if data science can be automated, but if, let say 80 % of it can be automated. I would argue for a big yes.
Hence, even after the economy is open, it would be hard to get those people back to work until the "enhanced" benefits run out.
I.e. I expect to see a lot of restaurants/hotels open but with no employees. Or, they will be employees, but their wages will have to be increased by 50%, which would cause runaway inflation.
If this happens, this might lead to spike in automation (AI).
So based on that, even if the economy is open, I failed to see what would incentivize people to return to work.
In addition, for your customers, the UI IS the product. Especially for the CxO level. So most of the "go/no go" decisions will be done based on the quality of the demo UI.
For a startup, you want wide and somewhat shallow experience. Since everything is an experiment, you want to crank out features and see what sticks.
I.e. you want a programmer that can do some UI / some app logic / some data science / some database / some infra/ops, (Not unlike a special forces person).
If you come from a big company, most roles are very specific / very deep - for example, a person works for 2 years on a file system kernel.
First of all, you can close the studio. I am not sure what is the legal structure of the company and if there are any personal guarantees, but this is not a regular situation. I.e. if you do not have customers, how do you suppose to pay your rent.
If not:
1. Try to offer 1-1 remote sessions. You might get more than 5$.
2. Go behind your regular customers. For example, try to go after companies that are operational and would like to offer Yoga to their overwork staff - e.g. warehouse employees, Hospitals etc. They might be able to charge your lessons on some sort of relief fund.
3. Ask companies that move remotely if they can introduce remote Yoga as part of their remote schedule. I.e. as a way to keep the team engaged. Companies are usually more willing to pay than consumers.
The economy is flooded with cash (3T fiscal injection + 10T form the FED). The FED basically bought most of the outstanding treasuries and injected cash into the system).
So once the virus is over (mainly due to massive testing), there might be inflation, which means that equity will be much pricier (not cheaper) in nominal terms.
I.e. if the economy was heated prior to the virus, it would be 2X heated after the virus.
In addition, there might be a shift of supply chains from china, back into the US, which means more jobs and more software.
However, we first need to get over the virus.
If mass testing is adopted, it would surface the asymptomatic carriers.
In addition, it would allow the creation of "clean" zones (.e.g on flights/buses) where you would need to prove that you are clean.
So, once the virus cannot "hide", and R0 is bought down to below 1, you would actually see the rate of infection crash to zero (I.e. the exponential rate works both ways).
To sum up, as long as there is no vaccine, it all depend on mass testing.
Once we have massive testing, we must create clean zones in order to allow the economy to recover.
If we do not create "clean zones", the virus will return as soon as you lift the quarantine.
You cannot stop this virus unless you somehow tag who is "clean" and who is "infected"
I.e. a solution to this issue would require:
1) Massive testing of everyone. This can be done via a mobile lab or "at home" kit.
2) Issue a clean certificate
3) Division of all areas to "clean" (you can prove that you have been tested) and "infected". For example, you can board a flight or a train only if you are "clean".
In general, the initial test should be done in parallel and hopefully at the same time.
Once this test is done, all the "clean" certificate can be issued and the quarantine can end.
I personally cannot see any other way. I.e. even if you end the quartine after X weeks, the virus will start again.
So the race is either to massive testing or to a vaccine. I think that reaching massive testing is easier.
Hence since there is no way to prove if one is infected, the government assumes that everyone is infected, thus halting economic activity.
The solution will probably be:
1) Massive testing - I could see "at home" test kit. 2) Some sort of certificate of the day of the test and the results. 3) Some sort of division between clean areas (where only non-infected people reside). For example, a flight must certify that all the passengers are clean.
The other option is to have the virus return every X month until there is a vaccine, or until 90-80% of the population is infected.
The core of the issue is that there is neither a vaccine or a test, combined with asymptomatic carriers. Hence, the only way to combat this virus is a quarantine for the whole population, which causes severe economic damage.
Once you solve either of these issues (e.g. South Korea), there might be a fast recovery. Did South Korea enter into a depression?
In any crisis, it is easy to assume that the last crisis results will occur, but everything depends on the context.
The other thing which is critical is creating a COVID-19 home testing kit or any sort of testing automation / automatic testing pipeline.
The only way out of this situation is extensive testing, but from reading around, the current testing time (from mouth to result) is around 4 days.
I think that testing will be an integral part of daily life until we have the vaccine.
I can also envision that verifying that you are not a carrier, will be part of a day to day procedure.
Also, price is always leading the news, not the other way around.
And, a lot of fund managers (who already entered the long positions) will go on TV and start suggesting their own version on how to handle the crisis and why this is an opportunity of a lifetime. Assume that everyone is talking the book.
So a bear market is lower lows/lower highs. The SP is still in this bear market. As long as this bear trend continues, do not enter.
However, there will be one day in the future when the market will reach capitulation. The news will be so bad and panic so high, that the market will fall hard, but then you would see very high volume of buyer, and the market would end positively for the day.
At that day you can start to gradually enter.