HN user

kaib

660 karma

Founder and CEO of Airstone Labs. We work on supercomputers. If you are interested in hard systems engineering problems get in touch, we are hiring in San Francisco. @KaiBackman on twitter. kai@airstone.io

Posts3
Comments163
View on HN

More specifically when the user clicks on a specific point on the planet and moves the mouse cursor you want that point to still be under the cursor when the mouse button is released. You can extend this construct outside the central sphere into a camera control that works across the screen and feels very intuitive at any zoom level.

Google "sgi trackball" for some interesting code to point the way.

YMMV. My first memories start as a trickle before the first year growing to a torrent starting from 12 months and forward. My sister hardly remembers anything before the age of 10. :-)

I'm actually interested in a bit more detail as well.

I've seen a few different fabrics in production (including proprietary non-public ones) and while I'm not sure IB is the answer Ethernet certainly has quite a few performance limitations.

Leaving Go 12 years ago

Does Ken count as a Systems C programmer?

Most of the systems C programmers I know, who admittedly work on large distributed systems, like Go.

Toward Go 1.3 12 years ago

It works fine. When I did the initial implementation of the Go ARM compiler I mostly used my phone as a debug environment. Android was probably the only Linux distro where the compiler worked when 1.0 was released.

Having been raised in a cold and isolated country (Finland) I think you incorrectly assume that means a lot of time spent indoors. This can be true, but does not have to.

I do think you are on to something with the idea of pods, most farmhouses where I come from are arranged in a number of small separate buildings.

Go 1.2 is released 13 years ago

Why wouldn't split stacks work on SPARC, is there something about the architecture that makes it impossible?

Four years of Go 13 years ago

That is almost doable, but would be exceptional unless your name is Rob, Robert or Ken. I think I could almost claim 9 years at that point, conveniently ignoring the fact I stopped coding a few years back.. :-)

I'll counter your anecdotal evidence with my anecdotal evidence. I know a few systems guys who have three or more decades of experience with statically typed languages who prefer Go. :-)

My impression is that static linking is undervalued by developers who have not been responsible for the production deployment of binaries to larger (hundreds or hundreds of thousands of systems). When debugging production systems at that scale you want to minimize the number of moving parts given the problem is so hard to begin with.

ABI compatibility just postulates that the bytes entering and leaving the interface will be marshalled the same way. However, the semantics of the function must by definition always change when you change the implementation.

You are arguing that these semantic changes are sometimes inconsequential, like for security updates. Based on my personal experience I think a more conservative view is appropriate and any changes to the dependencies of a program should be tested before large scale deployment.

Extrapolating from your use of "they" and the implied "us" I'm guessing that you haven't worked in the role of a CEO yourself. Taking that responsibility for a few years might not increase your respect for individuals but it will increase your respect for the difficulty of the job.

One important observation is that the interconnect out of the Phi cards is note very good. While the interconnect inside Phi is pretty good the bandwidth to communicate with other computing facilities is through the PCIe bus. This places the Phi cards close to GPU's very good at graphics and a few specialized problems like protein folding but not supercomputers that can tackle hard problems like sparse matrix solving.

San Francisco 13 years ago

The comment on schools is incorrect. The school system changed a year back and now you have a very high chance of making it to your neighborhood school. This is both from reading the system description and first hand accounts from friends with kids going to Kindergarden.

Noe Valley is pretty nice and it's hard to say if it's more expensive than say Mountain View in the south.

The short answer is no, we only started selling the assets three weeks ago, a few months after announcing the shutdown.

The longer answer is that selling part of your business while continuing in a business that is in any way related is usually very challenging from a legal standpoint. The lawyers call it the "spaghetti effect", basically the IP tends to be really tangled. We had serendipitous developments a few weeks back in how Airstone is built which let us create a hermetic barrier between the Airstone and Tinkercad IP's. Before that happened we had no way of selling the IP without taking an unacceptable risk.

Autodesk engineering will take over all Tinkercad development. We are working on transitioning everything over, Autodesk has built a very strong and dedicated team to support Tinkercad.

Airstone development will continue as before.

Owning less of a much larger pie does seems sensible from a monetary standpoint. It seems like you are optimizing for percentage ownership of the company instead of actual value of your stake?