sseveran

Its cool. I will buy. And you will buy in 12 months.

This is going to cause some disruptions. What are the alternatives out there to bitnami charts?

They had an engineer as CEO. That one did not work out.

https://en.wikipedia.org/wiki/Dennis_Muilenburg

Some context can be found here: https://www.nbcconnecticut.com/news/local/ccsu-professor-sus...

And the US built a global network of microphones to track them called SOSUS.

https://en.wikipedia.org/wiki/SOSUS

Oh, I don't blame him at all. A commercial airliner is a decade long project to undertake. However its clear he was not the leader to lead a company through a crisis.

The CEO that Boeing fired at the end of 2019 was an aerospace engineer who had been with the company since 1985.

https://en.wikipedia.org/wiki/Dennis_Muilenburg

A quick look at their 10Q does not show that their Energy generation and storage business is extremely profitable. It eeks out a small profit before operating expenses. That may change in the future. It is also a small business relative to their car business.

https://www.sec.gov/Archives/edgar/data/1318605/000156459020...

There is at least one YC company that would beg to differ with respect to airlines. Boom Technology.

https://boomsupersonic.com

If those kind of risks were imposed on banks they would only work with businesses they knew very well or not at all. As it is there were significant questions from the banks side on what they would be liable for if the loan was ultimately found to be invalid, or fraud.

There is nothing as permanent as a temporary solution.

1mbps is 324 GB in a month.

Stepping in between buyers and sellers is called market making. They smooth supply and demand across time.

Also I don't define front running that way, FINRA does.

https://www.finra.org/rules-guidance/rulebooks/finra-rules/5...

It's not basically front running. Front running is specifically when a broker gets your order and sits on it until either they or another client can trade in front of it before the market moves. Market makers want to interact with retail orders since they typically don't have size and in aggregate are either non-directional or predictably directional. This reduces the market makers risk of adverse selection.

I agree with this, although I think it was a conscious and deliberate choice with TF 2.0. We have given up on TF for all future work which is sad since I really appreciate a number of the pieces that surround the core. I think they made a choice to emphasize the support of already developed models and make the experience great for novices will be a decision that they will come to regret. We found so many issues when we tried to port some of our existing models to TF 2.0. The sad part was that there were GitHub issues for all of them.

Personally I think Tensorflow has already lost and we just need to let it play out over the next few years. One interesting wrinkle is that since Trax, Jax and Flax utilize pieces of Tensorflow the TF team can probably claim good internal adoption numbers depending on how they count.

I agree it will probably be use case based, but maybe not. It could be that at least one state bans using software to recognize faces. I believe San Francisco had to amend its facial recognition ban to allow the use of iPhones.

I also think a lot of people calling for a ban have not really thought through what would be banned. Do we ban models trained with triplet loss over any biometric data? Is it just specific entities that are banned from using it? Is it ok for Apple to build me an album of pictures of my wife and I? How do we define face recognition?

This is an incredibly complex and nuanced topic and it seems like many people discussing it just wave their hands without considering what the ramifications of any overly broad definition might lead to.

https://www.businessinsider.com/san-francisco-amended-its-fa...

The Tuskegee syphilis experiment was run by US Public Health Service which is currently part of the HHS. The FDA was founded in 1906, although it has earlier origins, which is well before the study was begun.

I mean that is a pretty strong statement. It's not like it was really possible to criminally charge Tony Hayward with being at fault for a faulty blow out preventer design, even though a number of people died. In some ways we were probably less safe with him fired, given that one of his objectives was to develop a much stronger engineering centric culture at BP.

In Brazil Vale and a number of its executives (along with the European dam inspector) are expected to be indicted. It will like be about falsifying documents related to dam safety.

https://www.reuters.com/article/us-vale-sa-disaster-exclusiv...

Proving that the CEO is the one that should go to jail, while satisfying, is actually quite difficult. While I would probably agree with one piece of your sentiment, senior executives are often able to escape prosecution for things that are clearly illegal (looking at you HSBC), simply trying to lockup CEOs I don't think is the right way to go.

VW's former CEO isn't being charged with killing anyone one. He has been charged with committing fraud by knowingly lying to regulators.

Dealing with the corrections is just another product. When I worked with IBES data, which collects all these estimates, you needed to buy a package called something like "As Was" to recreate any arbitrary point in time estimate.

In fact most of these articles neglect to point out why the tax rates are what they are or offer specific policy prescriptions. I would at least respect them if they said something like "let's get rid of the R&D tax credit." SarbOx made stock grants an expense that is recognized so we could undo that which would greatly increase the net income that is taxable by many of these companies given their reliance on stock based compensation.

Without discussing which provisions (and they are actual laws, not loopholes, because many/most are intentionally there) we should be eliminating this is all just a lot of handwringing.

Most teachers work for local/state governments not unstoppable modern corporations.

So we have been a TensorFlow shop since well before 1.0. We are keeping tabs on PyTorch with an eye on moving in that direction as certain things get filled, notably serving and something that is really comparable to tf.data.

The fact that our team in particular is looking at moving is probably a disaster for TF since a number of the criticisms typically leveled at it we didn't see as issues. We have a couple of really strong people with haskell backgrounds so static graphs and laziness we didn't find to be problematic.

However the embrace of Keras in 2.0 has left us dumbfounded. On one hand having a consistent layer interface is nice. On the other hand having a base class for the loss function that is not sufficiently general, the fact that all non-toy models we build seem to need model subclassing and a custom training loop with GradientTape and the number of issues we ran into while trying to port a couple of models has led me to conclude that the release was not ready. So while we like the tools around the model (tf.data, tensorboard, serving, tfx, etc...) building actual models I think has gotten worse.

Now my opinions on PyTorch are not from shipping production models but mostly porting to TF and keeping tabs on what they are doing. PyTorch also makes it easy to define reusable units. It does not try and expose a higher level interface that requires a significant investment in learning to express complex or unusual models. It seems a bit less opinionated on what the user should do.

A couple of other notes, PyTorch is being used inside Google for research I think. They have now written several papers (including one with Jeff Dean as 5th author) that have had their code released in PyTorch. PyTorch I think (it might already have) will end up with better governance but I would be interested in others opinions. They have at least one person listed under the project maintainers who does not work for FaceBook. A reason for adopting PyTorch may be that one company just does not decide to radically change the project to fit their view of the world.

This last bit is purely conjecture. PyTorch I think has already won over TF and it is going to take a couple of years for it to play out. If I had to bet today I would bet that PyTorch will become the dominate framework for both research and production. Of course something could happen to derail that but if things continue on their current trajectories I think its inevitable.

You can buy insurance for this today.

https://www.travelers.com/business-insurance/oil-gas/midstre...

Why would a patent troll sue a rice farmer with no money in a developing country?

It's irrelevant. You need the CUDA SDK and the CuDNN SDK both of which have license agreements.

Edit: You may also need the official Nvidia driver. I have never run anything with the OSS drivers.

Its because NVIDIA requires Teslas in the cloud, which while more powerful than 2080 ti's, cost roughly 6-8x as much.

I am not sure it's just IT competence. If this was the worst thing going on in Baltimore it would be doing much better as a city, but its not.

The US was most definitely a nation before the Civil War. And the Civil War was waged to prevent people that identified with each other from founding their own nation.

I completely agree the website owner has the right to run whatever they want, so thanks for saying that as its something that too often gets overlooked.

At this point I basically refuse to use things with recaptcha or the stupid little cloud flare dots. I just close the tab and move on. I am just so tired of little dots, storefronts, cars, etc...