Hardware Podcasts

AI for climate: a real world test

Catalyst · 15 Jun 2023 · 1h 19m

AI in practice: one real use of AI described

What this episode covers

Duncan Campbell, vice president at Scale Microgrids, and Seyed Madaeni, CEO of Verse, test whether a non-coder can use ChatGPT-4 to build battery dispatch software. Campbell iteratively prompts the model to create a Python optimizer for a 25MW battery in the CAISO market, successfully generating code that retrieves historical prices and calculates optimal charge/discharge schedules. Madaeni evaluates the resulting code, noting its clarity and optimization logic while highlighting missing complexity like ancillary services and forecasting. The episode demonstrates that AI can significantly accelerate the development of analytical tools for energy markets, reducing a task that would take engineers weeks to hours for a domain expert.

AI in practice

Where this conversation describes AI being used to build or deliver a product. Part of the AI in practice tracker.

Engineering & simulationChatGPT

Duncan Campbell iteratively prompted ChatGPT to build a battery dispatch script from scratch, teaching it constraints and logic through detailed natural language descriptions and debugging errors, resulting in a functional Python notebook.

“I actually tried to give some examples too, with some like basic, just arithmetic to demonstrate what's at work and let it interpret how to turn arithmetic into algebra. Yeah, I tried to really sort of describe how it all should work up front.”

Duncan Campbell · Hear it at 19:56

Most relevant to

early-stage companies

Topics

Mentioned

Companies
Verse, Fluence, Tesla
Tools
ChatGPT, Grid Status
Also
CAISO

Related episodes from other shows

Other shows on the same ground, closest first.

  • The Next Byte Podcast · 23m

    151. Battery Chemistry GPT

    Hosts Daniel and For Bode discuss a University of Michigan project developing a GPT-style AI to design battery electrolytes, aiming to lower EV costs and improve energy density. The team utilizes the Polaris supercomputer and partners with Deep Forest Sciences, adapting drug discovery methods to material science to predict viable electrode-electrolyte pairings. They highlight how this AI-driven approach, combined with automated robotics for testing, could accelerate the discovery of sustainable, affordable battery chemistries for the construction and consumer EV markets.

    Batteries & low-power designAI & machine learning
    7 chapters

  • The Amp Hour · 1h 13m

    Like a steam engine in your house

    Chris Gammell details his new 13kW solar and Tesla Powerwall 3 installation, discussing the economics of virtual power plants and peak pricing. Dave Jones analyzes Australia's energy market, highlighting negative wholesale prices and the high cost of peak power generation. The hosts also discuss the impact of new US tariffs on the global electronics supply chain and the challenges of designing with tiny 0.35mm pitch BGA components.

    Energy & climate techPolicy, trade & tariffs
    9 chapters

  • Oxide and Friends · 1h 39m

    Does a GPT future need software engineers?

    Bryan Cantrill, Adam Leventhal, and guests Keith Adams, Ashley Williams, and Jacob discuss the implications of GPT-4 and large language models for software engineering. They analyze the tool's strengths in code generation and debugging versus its hallucinations in niche domains like DTrace, while debating whether AI will replace engineers or serve as a productivity multiplier that lowers barriers to entry for new developers.

    AI & machine learningCareers & education
    10 chapters

  • Hardware to Save a Planet · 31m

    Powering the AI Era Sustainably: Breakthrough Climate Hardware Tech from 8 Innovators

    Host Dylan Garrett presents a year-end recap of eight innovators building climate hardware to support the massive energy demands of AI data centers. The episode features insights from Arch Rao of Span on managing grid capacity, Anders Kousgaard of Blue World Technologies on hydrogen fuel cell complexity, and Ina Braverman of Eco Wave Power on harvesting ocean wave energy. It also covers Ted McElveen of Vern on replacing diesel backup generators with hydrogen, Holly Jane of Antora Energy on thermal batteries, Curtis Van Wallahim of HydroStore on compressed air storage, Sandeep Nijwan of Electra on low-carbon iron, and Greg Williams of Sublime Systems on decarbonized cement.

    Energy & climate techScaling production
    9 chapters

  • Robot Talk · 19m

    Episode Twenty-Four: Gopal Ramchurn

    Gopal Ramchurn, Professor of AI at the University of Southampton and Co-CEO of Empati Ltd, discusses the design of responsible AI for socio-technical applications like energy systems and disaster management. He explains how Empati uses AI to optimize green hydrogen production from intermittent renewable sources, addressing the challenge of ensuring the electricity input is guaranteed green. Ramchurn also details his role directing the UKRI Trustworthy Autonomous Systems (TAS) hub, a £33 million program focused on creating best practices for trustworthy autonomous systems through interdisciplinary collaboration.

    AI & machine learningEnergy & climate tech
    5 chapters

The summary, chapters and topics are generated from the episode’s audio by this directory, not written by Catalyst. The episode itself belongs to its publisher.