The Winter Holidays are marketed as a time for celebration and gifting. However, for me, they always had a different sort of significance: an opportunity for deeper introspection. Taking this opportunity, I will write down my priorities for 2020.
Publishing this serves as checkpoint, proof-of-integrity and reminder-of-self-accountability for the data presented henceforth.
P1. Business Objectives
- Improve the interfaces by which humans and machines learn from each other. At Knosis.ai, we will continue to expand our machine learning and our data tagging pipelines. With the first customers secured in Europe, our attention will be focused on international expansion, both of the supply chains (Eastern Europe, India, maybe Africa) and of the demand pools (Western Europe, US).
- Understanding the importance of privacy-by-design, at Envisage.ai we are currently working on TrustBridge, a crypto-biometric mechanism for Know Your Customer (KYC) strong authentication that does not, at ANY point, share ANY biometric data with the data processor or operator. Thanks to modern cryptography which is built into both iOS and Android phones, we are able to insulate the “guarantee of correct and dutiful processing” from the “custodial ownership and control of the sensitive data data” (pardon the legalese).
- Stronger support for “beefing up the stack” of local IT partners. There are a lot of companies in Romania who dream of “artificial intelligence” or “smart IoTx” projects, but still use pen and paper for keeping track of work order and expenses. Together with our partners at Jiratech, we are open to sharing our knowledge with committed to sharing our local partners, in regards to development automation, machine learning, non-linear data models and holistic development flows.
- Going back to the roots. Literally. As the proud new partner in the Enten venture, I will be having the opportunity to unleash all the integration and data mining pattern (in short, “geek-ness”) from the laboratory in the field, so as to increase agricultural yield, reduce pesticide reliance. At Enten, we use an array of tech (proprietarily-design IoT points, in-field observation, soon satellite imagery) to help farmers plan, seed, work and saw more reliably, with less cost and less exposure.
- An overarching objective in the three business is increase the pace of our intellectual property strategy, by filing 5-6 patents that we have already prepared involving human-to-machine learning, distributed ledgers, eventual audit, data and abstraction encoding automation, visual machine learning and open, trusted, computation marketplaces. We are open to discussing partnerships in non-European jurisdictions.
P2. Open Standards and Market Formation (Deep Vision; non-profit)
When the team and I started to work on DeepVISS we considered it to be just a technical standard that helps us integrate visual machine learning and computer vision components, interfaces and data faster. We made this contribution of ours open source on GitHub.
We decided to rethink DeepVISS as just a component a larger whole: Deep Vision, a non-profit task force focused on informing, enabling and encouraging consumers, businesses and institutions alike to be both proactive and cautious regarding the potential use and abuse of recklessly-deployed technology, cryptography, excessive automation. Of course, DeepVISS (Deep Vision Interoperability Specification Standard) and the commitment to contribute to open source shall remain one of the core tenets of Deep Vision, in its entirety.
This once again means that we, as professionals, have to get out of the laboratory, as one of the problems we discovered is not lack of technology, but lack of business and consumer education around the opportunities, implication and pitfalls of ever-disruptive technology. For this avail, we have decided to participate by publishing a series a white papers and academic papers in the first quarter of Q2.
P3. Community Involvement (Deep Vision; non-profit)
Together with fellow professionals (IT, business, law, finance) and the non-profit community, we have worked on a roadmap of technology solutions and recommendations for government and public sector, focused around “compliance-by-design” – the principle of baking in legal and regulatory compliance standard in the fundamental way application, services and technology in general are built.
Deep Vision will publish its first roadmap of recommendations and initiatives by 31st of December 2019, so that they are available from the first second of the new decade.
P4. Personal goals
Managing two start-ups is the worst time to write a book. So Murphy’s laws state that it’s also when you’ll feel the most inspired.
For this reason, in the past year, some of my evenings/nights and weekends have been spent writing, sketching and doodling away at a book which I eventually titled:
Embodied computational calculus
(lat. Calculis computus incorpore )
(ro. Analiza calculabilității încorporate)
The focus of the book is the search for a model for inter-disciplinary reconciliation between the tenets of computer science (ak. the Turing world), the world of more-or-less theoretical physics (string theory, holographic theory, M-brane theory and so on) and the real-world implications on and perception of living things (us, dogs, microbes, aliens, AI). By no means do I or will ever claim expertise on all such fields. Quite the contrary: the purpose of this book is to outline a speculative-at-best proposal and to draw-attention to the interdisciplinary crevasses that prompt further attention and investigation to be required.
Since we are in the digital age, I don’t think publishing a book has to mean one big moment when you press the big “PUBLISH” button. I think in the modern world, writing is a distributed, collaborative process. So the whole book will be designed as an interactive experience, where reader also get to be writers or reviewers, depending on how soon they join.
If you are interested in participating in this journey, take 2 minutes to let me know.
This is my plan for 2020, open-sourced.
Comments and pull requests are welcome :)