Guillaume Bouchard is the Co-Founder and CEO of Checkstep. He is also a ML researcher and trust and safety expert. He previously co-founded Bloomsbury AI to reduce the prevalence of misinformation using AI, which he sold to Facebook in 2018. Now he spends his time building content moderation systems to reduce harm and hate online.
Join us today as we cover a range of fascinating topics like doing ML research inside a French castle, building and selling AI start-ups, relocating with a wife and 3 kids from France to London, the distribution of human genes through dating apps, and the opportunities and risks that come with the explosion of generative AI.
Follow us:
Guillaume Bouchard - https://www.linkedin.com/in/guillaume-bouchard-2494882/
Alex Alexandrescu - https://www.linkedin.com/in/alex-alex
Sofiane Bébert - https://www.linkedin.com/in/sofiane-b
References from episode:
Andrew Ng’s Paper about Discriminative vs. Generative models – https://ai.stanford.edu/~ang/papers/nips01-discriminativegenerative.pdf
Ilya Sutskever’s Paper ‘Sequence to Sequence Learning with Neural Networks – https://arxiv.org/pdf/1409.3215.pdf
What is federated learning? – [https://www.v7labs.com/blog/federated-learning-guide#:~:text=Federated learning (often referred to,model locally%2C increasing data privacy](https://www.v7labs.com/blog/federated-learning-guide#:~:text=Federated learning (often referred to,model locally%2C increasing data privacy).
‘The Age of Infinite Misinformation Has Arrived’, Gary Marcus –https://www.theatlantic.com/technology/archive/2023/03/ai-chatbots-large-language-model-misinformation/673376/
Timestamps:
[00:00 – 08:21] Early years of generative AI - ChatGPT’s brother from another mother.
[08:21 – 14:08] Building generative AI for knowledge bases.
[14:08 – 23:58] Growing and recruiting at Bloomsbury AI.
[23:58 – 31:22] Fundraising and building a content moderation system at CheckStep.
[31:22 – 33:13] Is scale all we need to further develop generative AI?
[33:13 – 35:07] Using AI to improve research output.
[35:07 – 40:40] How to fight against ‘The Age of Infinite Misinformation’.
[40:40 – 45:10] Can we remove bias? What are the implications?
[45:10 – 47:04] The role of content distribution platforms and how they should be managed.
[47:04 – 49:30] Breaking into Data Science.
