I just started testing Keybase filesystem. Looks promising so far. Everything encrypted, 250GB for free (for now). Sharing folders with my coworker works seamless.
The advantage over Dropbox is that it does not take any space on your harddrive. Also everything is encrypted by default.
A disadvantage is that your data is not available without internet, since it works like a network drive.
Running a Bash kernel in Jupyter Notebook might be convenient if you are running your Jupyter server on a remote machine. So you don't have to SSH into the server for some quick changes, you can stay in your browser. I haven't used it myself as I only use it locally.
They took the notebook part from IPython and made it language agnostic, this is now Jupyter. You can use Jupyter Notebook together with many languages (Jupyter Kernels) such as R, Julia, Bash and many more in addition to Python.
I think the clumping is more about how many consecutive same numbers there are in the array.
For example, a naive me would generate: [3,4,5,1,2,4,2,1,2,6]
whereas a true random distribution might generate: [3,4,4,4,4,1,2,2,2].
When you generate 6 million samples you would indeed have about 1 million per number. However you would still have subsequences of the same number.
When developing for microcontrollers and for example my leds aren't working, I like to define a place in memory where I write values to depending on the place in code. Then just read out that memory address with OpenOCD (jtag of swd).
I've had the same experiences with the C18 compiler. Using that compiler, building a simple 6-channel PWM application becomes quite difficult. For every interrupt, the compiler would copy the entire (call)-stack to a different memory location, then enter the interrupt routine and finally copy the entire stack back, resulting in latencies of ±100 cycles before and after the interrupt.
So, to get rid of the timing jitter, we had to fire the timers early, and then wait for the final timer counts in the ISR.
Later, we switched to the NXP LPC platform using only open source tools: GCC ARM Embedded toolchain, OpenOCD + gdb for debugging and vim, make as 'IDE'. What a relief.
That is better given one is more comfortable with mechanical systems. Since my field is EE, I tend to do the opposite: translate a problem to the electrical domain to be more intuitive about it.
You can use Bond graphs[1] to easily translate between different domains (Mechanical translation,Mechanical rotation, Electromagnetic, Hydraulic and Thermic in a more limited way). This technique generalises force and voltage to "Effort" and velocity and current to "Flow".
Might be a good accessory to sell.