cjhanks

Recently it's failed to find functions by verbatim search name within single project searches. I am pretty sure it used to function better than this.

The bank was nearly fully capitalized with low risk investments. Everyone would have gotten their money back anyways, it would have just taken 5-10 years.

Imagine what would happen if startups could not access their capital? By the time funds were available for withdrawal, the businesses would be dead. The regulators then have to track every funder to every dollar to return money. It would take a decade for this to happen.

I didn't see catch this part, but having 10 notes over 4 octaves is not out of the question. If all of those notes are rhythmic, that could be a challenge. All of the fairly wide chords I saw could be played either with sustain or could be easily substituted with spread chords.

Pop pianists don't usually play what's exactly on the page.

Hopefully not too off topic. How do people monitor these BGP announcements? Is it possible for laymen to monitor such route rules?

Question. From my experience, GitHub uses the `github-amo` bot to fetch the data when the page is rendered. How did you prevent it from cacheing on GitHub's servers?

Wake up with the sun or when I am rested.

Eat eggs or whatever I have around with a lot of milk and a cup of coffee.

Take a shower.

Read until I don't want to anymore.

Go to work until I am hungry.

Eat something, usually around 1pm.

Take a walk.

Work a little more.

Go somewhere to play music or sit with friends.

Shower.

Go to bed when I'm tired.

Doesn't this mean that a conditional branch is added to all existing code which performs casting?

Some societies believe it is acceptable to judge people based on their appearance. Arguably, most people do it instinctively. It is just that in America it was deemed illegal. I know there's an epidemic of plastic surgery taking place in some Asian cultures - people getting cosmetic surgery so that they can be employable.

Right or wrong, that is how some societies are functioning.

Why is showing your child a professional boundary? Different phases of life necessitate different accomodations.

This person is advertising their priorities upfront. Startups might see it as a weakness. Established companies might see it as a sign of stability.

LinkedIn is a historical anomaly anyways. It's generally considered bad form to include any photo on a technical resume.

It probably helps that Julia and Fortran are both column major.

When interfacing with C languages, that can be a big penalty (depending on the complexity of the algorithm).

It will probably take two generations of developers to make this happen, but well worth it. When I was studying atmospheric science I tried to understand OpenWRF (weather research forecasting) and the barrier was too high... And I had programming experience, imagine those with none!

An open source project is only as useful as it is accessible to it's average developer base.

Somewhere there is a programmer that tried to fix a bug in 10+ year old legacy code (mso98win32client.dll) who has learned the bugs of legacy are features of today.

I really hope they do not do this. In and of itself it is not a problem, Ruby code is perfectly readable when the style is consistent.

Software developers are often stubborn. The main result of this change will be a mess of code which follows a mix of all the styles. That removes one of the greatest beauties of the Python ecosystem, that there is a "right way" to do things.

This presumes that the truth is knowable and you could know it.

Mathematics tells you what is not possible, but it tells you nothing about what exists. Every other discipline is separated from truth by bias in observation. And there are no perfect observers.

So, pushing "truth" is really pushing a specific perception.

That said.. you can fight against "false".

Ooh, I missed those. Support for C++17 is pretty major. Thanks. Perhaps my memory is fuzzy, I just remember the CUDA 9->10 switch having some significant (but not major) performance and feature changes.

It is possible to SIMD parallelize multiple linear problems when you can guarantee convergence of each subproblem and compute iterative bounds.

But sure, SIMD is no panacea. Especially when the latency between the vector instruction shares transistors with the previous instruction.

Does anyone understand why such minor upgrades resulted in a major version bump? Is this some sort of stability check point? Or some other versioning convention?

Teach them about cyclomatic complexity and then review their work in these terms. It gives them something concrete to target rather than trying to accomplish some ethereal notion of "proper abstraction".

Julia, R, and MATLAB. But yeah, it's not common. I end up reinventing this wheel about once a year.

What a beautiful conclusion for him on such a beautiful project. This man clearly loves the art of programming... so much, that he would give up power, just to pursue it again.

Thanks for the wonderful blog posts (over the years) and products.

Yeah, I think you got lucky. He sounds like a great boss. And, he was probably right.

I suppose from the perspective of someone implementing these models, yeah - it is linear, but it is not bijective. In a system with only one layer, that manifests as an alias (assuming the output dimensions are smaller). In a system with multiple layers of either `N->M` or `M->N`, those aliases tend to manifest as apparent "non-linearities".

So, I guess looking from the bottom up the system may look non-continuous and linear. But if you look from the top down, it would look continuous and non-linear.

Really, I am not sure which one is "true".

If you have ever opened up Excel or a similar program. One of the more useful options is to generate a regression line-fit on your data points.

One option is to specify a polynomial function, you can specify how many coefficients you want. One of the measurements is the mean-squared-error between the line-fit and the points.

You can add as many polynomial coefficients as you want, and you will be able to decrease the mean squared error. But the more polynomial's you choose, two things will be true:

1. The line-fit will be far more likely to go through the points.

2. At points in the line where there was no data, the line will less approximate the underlying physical reality.

That same mathematical property is what is relevant here. There is nothing inherently evil about non-linearity, when the non-linear math model properly maps to the physical reality. But when you over fit a line, many of the functional solutions may be completely wrong.

Most definitely insane and probably well dressed.

I agree with the notion that artificial intelligence is a graph of smaller problems, as is human perception.

The problem is a question of informational density. Biological systems are computationally very dense. Far more dense than the 4nm transistor fabrication available today, and with a far larger volume of size.

Consequentially, the computational capability of most AI systems is far lower than its biological equivalent. And as you find in most information finite discretization problems - the lower density information system will alias against the higher information system.

So, that means you will have a hierarchy/pipeline of computational stages - each aliasing reality. Eventually, you will find that your parameterization of each perceptual stage has a strange property. The size of each subsequent layer is important... but the relative computational space of each subsequent stage is even more important. Because mismatched stages results in nothing but numerical interference and noise.

And I think that is where we are today. The IQ of a krill shrimp.

We have also been watching these machine learning models for 6 months:

- increase the volatility in virtually every financial market they touched

- be exploited by adversarial learning networks to amplify funded propaganda as news

- use poorly contrived sentiment analysis to generate incomprehensibly meaningless news headlines

These non-linear "function approximators" have absolutely unpredictable and insane non-linear behavior where learned information was non-existent or sparse.

God help us all if one of these artificial intelligence devices is driving the road and sees a red stop sign that is a square, rather than a hexagon.

Offloading a hash function to a custom chip seems appropriate. I suspect they are deploying an entire specialized box, not just a custom processor.

Edit: Nevermind, I may have been thinking about the wrong end of the pipe.

He gets my money.

It seems like people have forgotten the point of Google+. If I recall correctly... there were too many disparate login systems which resulted in a fractured identity. There were various legal and technical reasons which they could not associate users amongst their different properties.

Then they sent out advertisements to get everyone on G+. And a huge number of people clicked, without even understand what they were creating. Then all of the accounts were merged into one, that had different policies.

And so, there was no longer a profit motivation for the product.

I have gone without any phone/computer/etc completely a few times, once for 4 months, once for 3 months, and then quite a few shorter periods.

The main thing I learned was that when using the computer, I was using it too much. Even at work as a programmer, it turns out you do not need to spend most of your time in front of a computer coding. A bit more thought outside of the digital world can allow your ideas to form clearer, and then it takes far less time to code it.

I am glad someone has at least fake-formalized the infectious nature of anxiety.