Josh Wong is the Founder and CEO of ThinkLabs AI. He beforehand labored at GE Vernova as a Normal Supervisor, Grid Orchestration. Josh Wong attended the College of Waterloo.
ThinkLabs AI is a specialised AI improvement and deployment firm. Its mission is to empower important industries and infrastructure with reliable AI geared toward attaining international power sustainability. The corporate is creating its flagship product, ThinkLabs Copilot, a digital assistant that comprehends the true world via proprietary physics-informed AI digital twins, offering a foundational mannequin for engineering methods.
Are you able to inform us extra concerning the imaginative and prescient behind ThinkLabs AI and what impressed its creation?
The imaginative and prescient behind ThinkLabs is a dependable, sustainable, and inexpensive power infrastructure powered by reliable AI. We perceive that the grid stays on the middle of the power transition. To decarbonize we should electrify. To impress we’d like the grid, and the grid actually should modernize. We consider the intersection of electrical energy methods engineering, AI, and cloud computing is the answer.
How does ThinkLabs AI differentiate itself from different AI startups within the grid administration sector?
The grid is complicated, and a lot in order that AI in itself can’t be taught concerning the complicated energy flows and operational processes that exist within the grid house. ThinkLabs mix the wealthy historical past and confidence of conventional energy methods engineering with AI, as reliable physics-informed AI, for confidence in scaled, automated inferencing and choice help for important infrastructure. It additionally takes greater than know-how, however an skilled staff that understands the nuances of the grid and the way utilities and regulators suppose. Our staff comes from the electrical energy methods house with confirmed observe document, together with founder Josh Wong who has offered his earlier firm Opus One Options to GE, and stands on the intersection of engineering, AI, and cloud computing.
What particular challenges in grid administration does ThinkLabs AI goal to unravel?
Automated analytics and suggestions for actual time situational consciousness throughout the grid, massive scale simulations, and steady studying and suggestions to mitigate grid constraints and optimize grid efficiency. Particular useful areas embrace:
- Insights – close to actual time state estimation of grid energy circulation, detecting congestions, voltage violations and the way capital property are literally utilized.
- Options – optimum dispatch suggestions, together with switching, grid units and DERs, for congestion aid, mitigate DERs interconnections, cut back losses, restore outages, and so on.
- Mannequin validation – validation and corrections in utility supply knowledge units for grid fashions, saving OpEx and growing operator confidence for grid operations.
- Operator’s Copilot – operator dispatch suggestions educated with grid physics, enterprise guidelines, customary procedures, and operational expertise, empowering workforce coaching and upskilling.
What’s the ThinkLabs Copilot, and the way does it improve grid planning and operations?
ThinkLabs Copilot is a digital assistant that that understands the true world with proprietary physics-informed AI digital twins that present a basis mannequin for engineering methods. It really works with utility planners and operators, to mannequin the grid into its “AI digital twin”, carry out excessive velocity and enormous scale analytics together with in close to actual time, and make suggestions on grid operations, plans, and designs.
Are you able to clarify what a physics-informed AI digital twin is and the way it advantages grid reliability?
AI by itself can’t be taught such a posh system because the grid with measurement knowledge solely. AI digital twins of the true world are educated by, work for, and work with engineering methods, therefore “physics-informed”. Coaching is finished utilizing massive quantities of artificial knowledge generated from engineering simulation. Conventional physics-only, impedance-based digital twins are deterministic and mathematically optimized, but challenged by knowledge high quality, excessive computing energy wanted, and gradual response time. Conversely, basic AI methods promise velocity, but sparse knowledge, hallucinations, and “black field” results issues mission important grid operations. A physics-informed AI digital twin gives clear and reliable analytics, resilient and sturdy in opposition to dangerous knowledge, quick response and motion appropriate for actual time operations, preparedness with massive pre-trained working situations, and a closed-loop, steady studying and enchancment course of.
How does ThinkLabs AI make sure the reliability and accuracy of its AI fashions in real-world situations?
Nature of physics-informed AI retains AI grounded, tied to the true world, and bounded by the true world. We additionally do steady studying and monitoring of mannequin efficiency.
What makes your AI know-how significantly fitted to coping with the complexities of contemporary electrical grids?
Being educated by decide engineering fashions, however dealing with the imperfect knowledge high quality of actual world operations. AI additionally convey a wealth of optimization and generative methods unmatched by conventional engineering arithmetic.
How does ThinkLabs AI’s know-how combine with current grid administration methods like ADMS and DERMS?
ThinkLabs combine as a Copilot with current ADMS, DERMS, and AEMS, which can stay as the basic communications and management platform, whereas ThinkLabs will layer on further intelligence and automation as just like a automobile’s driving help system.
What does the latest $5M seed funding imply for the way forward for ThinkLabs AI?
This seed funding has enabled us to spinoff and launch from GE, associate with a bunch of world class traders, put money into our staff and product, coming to market with our first business Copilot, and work with quite a lot of channel companions to convey this into the palms of our prospects. That is the primary foundational step to subsequent enlargement and scale.
How do you envision the function of AI evolving in grid administration and different important infrastructures?
We see grid administration and different important infrastructure as being more and more “AI first”, particularly with physics-informed AI. Open up far higher understanding, situational consciousness, and growing automation choice making and orchestration of important actions. But, at all times stay humble and reliable as AI, being true to the inspiration legal guidelines of physics and engineering design.
Thanks for the good interview, readers who want to be taught extra ought to go to ThinkLabs AI.