Tech entrepreneur, web commerce pioneer, AI & software engineer and researcher.
Longterm project past decade: fixing how people communicate online, addressing everything from fake news to collaboration on science and other issues. In stealth mode while sorting out how to compete with big tech's resources while starting solo. (note: now set aside to deal with impact of new AI wave).
History of seeing tech trends before others (often too early), examples:
Founded Cyberspace Development in 1993, the first company to do ecommerce on the web at Marketplace.com, selling an ebook online months before Amazon was founded, providing stores and mailing lists for 3rd party vendors and teaching the public about ecommerce at events. Bootstrapped without the funding to wait for the market to grow.
In 1983 built framework to make GUI overlapping window interfaces on a PC before Microsoft Windows or the Mac came out, and created an object oriented C to do this before C++ or Objective-C were released to the public. Created for internal use in Carnegie-Mellon University funded startup MindBank. CEO declined to pursue it as a separate product. I was tech lead for the company and built a courseware authoring tool to create interactive computer with video disk based courses with a team that grew to 6 people. Microsoft's initial Windows had tiled windows due to technical challenges. They didn't manage on the PC hardware of that era to match our overlapping windows until December 1987.
Created TIA in 1994, a product for Cyberspace Development that helped bring people onto the net in the dialup era. It received glowing reviews, Internet World called it “Best new break-the-paradigm Internet software” and Wired "TIA is as close to magic as a graphically deprived net.user can get", and Internet Starter Kit books "TIA should seriously shake up the industry", as that "glorious creation" did, leading a user written FAQ to say:"IMHO, this is the best $25 I ever spent" and "equally if not more important was his excellent support of his customers". It was oddly even featured in a futurist publication in 1995 alongside AI, digital cash and nanotechnology.The net was far smaller then, a similar share of net users today would be several 10s of millions of people. (TIA Wikipedia page). [tech note: it allowed a unix shell dialup account to provide a home internet connection via slip or ppp using a new implementation of a tcp/ip stack that could run on unprivileged shell].
In 1991 prototyped a social media tool like Facebook for use on an intranet and predicted the eventual rise of problems with misinformation, filter bubbles and fake news if social media wasn’t done right, decades before it happened. Independently conceived and explored the concept of collaborative filtering and recommendation frameworks for when everyone was online, including ideas still not implemented.
In 2000 pitched a social media idea to venture capitalists well before social media existed, but it wasn’t funded due to the dot com crash leading to VC's misguided notions at the time that internet advertising was dead and that advertisers would have a problem with user generated content. Important aspects of it still don't exist, which has led to some of the negative impacts of social media on society. The system architecture planned a cloud framework using non-relational DBs when those weren't yet common, which also prompted VC resistance for not using RDBMS and big individual servers as was the status quo then.
In 1993 pitched a commercial internet browser several months before any existed, with extension language (not funded). Also independently conceived and explored ideas for affiliate and viral marketing, and varied other aspects of what the net would become in the future. Lack of funding resources led to not pursuing a stack of ideas that eventually became major sites like Wikipedia (though with approaches to deal with things like information credibility that may have prevented some of the flaws in Wikipedia and social media).
In 1991 before the internet allowed commercial traffic, did a prototype of an online newspaper with digital copy from the local paper. Included "books in print" as the likely "killer app" for online commerce, "Daily Me" personalization features, Google Maps-like features, hypertext with backlinks and user generated content from reviews to discussions. The plan was: local papers as dialup online services to be linked together in a national network since there weren't internet providers yet. Decided the industry would resist modernization, as they still do.
In 1983 designed a persistent object storage framework released as a commercial product, V-File by MindBank, over a decade before Hibernate & JDO popularized the concept, serving as software emulated virtual memory before PCs had the real thing.
Entrepreneur & Freelance Consultant since 1989.
Studied CS at Carnegie-Mellon University 1980-1983, undergrad. Expanded class projects to be more interesting, like a compiler-compiler ( parser generator and partly configurable code generator) instead of just a compiler, a syntax directed editor driven off a grammar description (hybrid with emacs-style text editing with customizable key bindings), graduate level AI, etc. Learned faster on my own.
Worked for the computer science department a summer, and a year full time, including AI at their Robotics Institute building a knowledge representation system using a distributed database approach.
Worked as an AI engineer on applied research at Martin Marietta in 1985-1989 seeking ways to apply Artificial Intelligence for internal use and for NASA and the Air Force:
Created a sophisticated AI system prototype with interactive GUI for use in strategic planning and analysis of future ground operations for space launch vehicles, allowing exploration of varied investments in resources and technology, logistics flows, impacts of new technology, cost and throughput. All the inputs and interactive graphs (drag points) interconnected with outputs like a hyper-spreadsheet with AI guidance using a GUI toolkit I built. Pitched to NASA and Airforce. MM received a NASA contract for a version of it. Responsible for problem selection/definition given "find ways to help the advanced launch systems group", system concept, functional specification, implementation, marketing demonstrations, and design of a larger system based on feedback on the prototype. Built on Symbolics LISP machine AI workstations.
Evaluated and reported on opportunities and estimated cost&benefits for the use of AI techniques in a MM manufacturing facility. Performed requirements analysis for scheduling system.
Sought internal support for my idea to do an intelligent CAD system to integrate all the stakeholders in the design process, e.g.: design for functionality, reliability, repair and maintenance, cost, etc. Didn't fit their current priorities.
Developed a fault prediction/fault isolation demonstration for the communications subsystem of the Flight Telerobotic Servicer being designed for NASA/GSFC.
Worked on creating new techniques for model based reasoning about physical systems.
As part of a NASA/MSFC contract to develop a prototype Space Station power system, developed a system to decide the priority of scheduled loads to be shed in the event of a power drop.
Most contracts were in the early 1990s before becoming an entrepreneur, samples:
Apple Computer; Cupertino, CA: April-May '91: In C++ on a Macintosh, built fieldable prototype of Eager, an AI system for automatically programming repetitive tasks by example based on watching everything a user does in an application and popping up when it sees something it may be able to help with. Based on an a LISP demo. Apple only needed it for one application, Hypercard, I generalized the program to make it easy to describe other applications using a knowledge base. The Apple researcher published about it, including a chapter in his book "Watch What I Do", which happened to document it as 15,000 lines, in 2 months.
InfoNow; Boulder, CO. The company developed disk based stores for MS-Windows that contained encrypted software customers could purchase and install immediately, which were included with some IBM and Gateway machines. 2 contracts with them:
June-July 1993: Developed code to allow users to "Test Drive" software stored encrypted on their disks or CD-ROM. When the test drive is activated with a key, patches to the OS make it appear as if the product is actually installed normally on their disk. The product remains encrypted except for a limited number of test drives. Microsoft expressed an interest in licensing the technology.
Nov. 1992 - March 1993: In C and assembler developed most of the GUI system to allow users to browse information about software they could purchase from the disk or CD-ROM store. Developed features to encrypt the program running the store and decrypt it on the fly, and other security features to prevent products being stolen.
McDonnell Douglas Communications; Denver, Co.: Jan. 1990- May 1990: Autogenerated C++ and SQL code for permanent object storage for distributed CAD app&more.
US West; Denver, Co. July 1989 - Jan. 1990: Developed decision support expert system in C++ and Clips rule language, tied to Ingres database with embedded SQL.
Those were just examples. Varied other contracts and attempts to found startups over the years that didn't fly, e.g. one derailed from funding by the financial crisis in 2008.
Summary: Full stack from concept to implementation of ideas, high level design to low level systems programming. Having built stacks, learning a new one isn't a challenge.
LinkedIn is here: rarely use it due to poor UX and thinking its a shame to support a dinosaur that should be disrupted, but put some info there since others do, and only public way to contact.
Some rough draft writings on AI and issues related to its impact on society on a Substack: Society and AI, including an essay how to address flaws in LLMs with hallucinations and difficulties with logical reasoning.
Boulder, Colorado based.