We are Alexander Nicholi and Charles Rosenbauer, two computer scientists based out of Raleigh, NC and Madison, Wisconsin, respectively. We share a deep background of research in programming language theory, high-performance computing, and general systems programming work.
We are looking to formally incorporate under the name Xion Megatrends.
We are creating a new personal computing platform from the ground up.
To this end, we have codenamed several product components:
On the hardware side, multiple credible efforts have proven that there is massive untapped potential in unorthodox hardware designs. REX Computing demonstrated this with funding from a Thiel Fellowship, but has not found success due to lack of a product. Adapteva also demonstrated this with $1.5M of funding from DARPA, but again ran into market issues due to how difficult it is to program these special chips.
This is why Anodyne focuses as much on software as on hardware: there is a clear desire to take advantage of more clever chip designs, but normally, the problem of backwards compatibility and software support remains.
On the topic of software, Anodyne will radically revisit what constitutes an operating system and vastly simplify the machine design and requirements of the SDK. In 1995, Linux was little more than a few thousand lines of assembly code; today it is 9.8 million lines of C code (12 million with comments). By focusing on the right kinds of complexity, Anodyne can provide all of the features without the bloat.
Simplifying one type of problem allows us far greater latitude to simplify the other, canceling out decades worth of unnecessary complexity that makes modern computers require thousands of developers to make product for. Simpler software allows for simpler hardware, and vice versa. Simplicity allows for more novel approaches to be taken without an unbearable cost. This is where we are at.
What’s more, our performance projections for the parallel compute chip are incredible: our best estimates point to a single 8,192-core chip being able to perform up to 4 quadrillion 8-bit operations per second (i.e. it is peta-scale). This is a general-purpose compute chip that is as efficient as a tensor processor – something major companies like Intel, ARM and AMD have not been able to come close to with their offerings.
This is why it is all the more important to build this now and follow it all the way through with software products that the general public can appreciate in the market. Both Microsoft and Adobe are heavily incumbent corporate players with very valuable products. They face no serious competition and will not be able to pivot in any reasonable timeframe to compete with this.
For now, we are raising a $700K seed round for the company of Charles and Alexander, to be called Xion Megatrends. The funding will pay for both of our salaries and all of the prototyping equipment and materials needed, up to and including FPGAs and MOSIS fabrication trials. We are looking to reside in either Austin or San Francisco depending on what is pragmatic, developing this product full-time for at least 1 year (maximum 2 years) while soliciting support for our Series A. Our prototype will vindicate our performance projections and our creative vision for the software.
Our Series A will be raised for the express purpose of taking our prototype to either Global Foundries or Taiwan Semiconductor, so that we may place an initial minimum order of real chips. We will finish a first edition of our Software Development Kit in the time this takes to procure and test. These chips will prove the performance numbers once and for all and galvanise Anodyne onto the world stage.
We don’t have a concrete plan by this stage, but generally we expect to be doing a lot of software development to complete Project Nocuous (the office suite) and Project Ciliary (the creative suite). Dogfooding of our hardware and SDK will occur at all levels while this is taking place, ensuring all of our core offerings are up to snuff for public use.
If this sounds like something you wish to invest in, you can email Alexander by visiting his website and reading the link out of the opening paragraphs. You can also reach out to him over DM on Twitter.