Kjetil Åmdal-Sævik

Data Science Manager & Machine Learning Engineer @ Oda - #machinelearning is both my job and my hobby!

[Kjetil Åmdal-Sævik @Keegil]

@Strumke Å ikke posisjonere datapersonvern og datadreven forbedring av offentlig helsevesen som en eller annen form for trade-off virker for meg litt uærlig. Mine data vil alltid være mer private jo færre som har og bruker dem, uansett hvor avidentifiserte de er.

[Kjetil Åmdal-Sævik @Keegil]

@Strumke "Vanlige folk" blir nok ikke eksponert for ubetimelige og edgy intellektuelle hot takes på Smittestopp fra Twitter, Medium og LinkedIn. Håper jeg...

[Kjetil Åmdal-Sævik @Keegil]

@Strumke Good point, and I agree! That's a much better gold standard, which will even hold in a hypothetical age of AGI. I understand your second point as well, although I guess I'm simply less concerned with it and very skeptical of it's realism. But I've been wrong before!

[Kjetil Åmdal-Sævik @Keegil]

@Strumke And yes, explainability will help contribute to trust (for quite flawed reasons, although that’s another discussion), but we have other proxies for trust which should be more than sufficient to make the world a better place with AI!

[Kjetil Åmdal-Sævik @Keegil]

@Strumke Again, I’m not saying that these aren’t important areas for research. We should keep trying to understand neural nets - not least because it will help us build even better algorithms and learn more about the world.

[Kjetil Åmdal-Sævik @Keegil]

@Strumke So as long as AI is overwhelmingly useful, I don’t see why it needs to be any different!

[Kjetil Åmdal-Sævik @Keegil]

@Strumke Now, I don’t believe a lot of people are saying that studying how and why drugs work is a wasteful pursuit, but suffice to say that humanity has a good tradition of trusting very valuable and important things - our lives - to things we collectively do not understand very well.

[Kjetil Åmdal-Sævik @Keegil]

@SilvijaSeres @BarMartland @tawaterhouse Haha, jeg vil nå si at «trafikkforståelse» er en bedre parallell, men jeg er jo enig i at dataforståelse ikke føles dekkende. Jeg liker digital skaperkraft, da! Datavisjon? Dataperspektiv? Vanskelig nøtt, @BarMartland...

[François Chollet @fchollet]

We haven't yet solved even 10% of the problems we could solve with existing AI/ML techniques. Even if new research were to deliver nothing from now on, there still wouldn't be another AI winter. AI/ML will keep on delivering for years to come.