Rethinking #AI for the future!
@PHenninger, a Partner and the UK Head of Connected #Tech and Global Head of #Digital Lighthouse at @KPMGUK , shares his AI fears and explores the future possibilities of AI beyond traditional use cases.
🎧: https://t.co/vaThlkEZlR
#TTLP
many recent chats w/ people thinking of moving to marseille led me to think about why Things That Grow Very Fast tend not to grow for very long, and how hype strangles good reasons for growth
(this generalises beyond marseille and cities)
https://t.co/46CB5DOGzu
Episode 4 with #data, #AI and #emergingtech expert @phenninger of @kpmguk is here! 🔥
🎧Tune in to #BeyondTransformation to get Paul's take on the role ESG data will play in reconciling the conflict between profit and a safer planet: https://t.co/B1V4x0jykV
Really fascinating framing of content moderation as a problem that has much broader application I think. Patterns and outcomes are the place to start….
In fact, once again, I challenge you to think about it this way: could you make your content moderation decisions even if you didnʻt understand the language they were being spoken in?
https://t.co/7WgSizOVl9
thinking clearly about risk may seem trivial but it's essential for good decisionmaking and good outcomes in situations of not-knowing.
read more about how to do that: https://t.co/l9G3fn6tdB
🙏 @jamescham@dglazkov@phenninger@gquaggiotto, and kai fong chng for comments.
Neural networks often pack many unrelated concepts into a single neuron – a puzzling phenomenon known as 'polysemanticity' which makes interpretability much more challenging. In our latest work, we build toy models where the origins of polysemanticity can be fully understood.
Lisa Dwan - Asking art not to offend is akin to asking to live in the world without sun, rain and wind for fear of their droughts, gusts and storms.
Actress and writer @Lisadwan on Salman Rushdie, friendship and art
https://t.co/md3BDO33fM
Some random musings on how the whole "AI scaling" debate. It is clear that neural nets can model any predictive distribution p(x_future|x_past) given enough parameters and data, since they are universal approximators;
@benedictevans The gaming side of Online Gambling has used AI to generate new products for a number of years (works because of the volume, speed, feedback AND ability to target specific demographics/behaviours).
Principal Component Analysis is one of the most fundamental techniques in data science.
Despite its simplicity, it has several equivalent forms that you might not have seen.
In this thread, we'll explore what PCA is really doing!
🧵 👇🏽
According to our 2021 #CEOoutlook Pulse Survey, driving the digitalisation of operations is a key focus for CEOs. Read @phenninger's blog as he discusses what businesses need to consider when assessing their own digital transformation journey: https://t.co/CiiB1IbreG