I’m Will Sentance, an ML/AI/computer science educator/founder – right now I'm in Davos, Switzerland, attending the World Economic Forum for the first time – it’s ‘insider’ as hell which is both fascinating and truly concerning
Proof here – https://imgur.com/a/davos-ama-0m9oNWK
It's full of people making decisions that affect everyone – v smart people like Andrew Ng (Google Brain founder), Yann LeCun (Meta Chief AI scientist) & lots of presidents/ceos
But there’s a total lack of transparency at these closed-door sessions – that’s why I asked the mods if it was cool to do an AMA here – and they very kindly said yes.
Here are a few key takeaways so far:
- AI is everywhere – it’s the central topic underpinning almost every discussion (and a blindness to other transformations happening right now)
- CMOs/CEOs (and people selling) say quite a lot of nonsense – it’s really hype train stuff from the fortune 100 "now we're doing agenticAI"
- The actual experts are both more skeptical and more insightful – Andrew Ng today was brilliant – tomorrow is Yossi Matias, Dava Newman
- OpenAI exec announced an “AI operator” (can handle general tasks) but defended their usual ‘narrative’- they’re so on-message every time w “AI is not a threat, just use our tools and you’ll feel great!”
I come from a family of public school teachers and I’m seeing how these tools are changing so much for them daily – but there’s no accountability for it – so I love getting to go in and find out what’s really happening (I did something similar for berlin global dialogue last year and had a more honest convo on reddit than there)
I’m here at Davos for the next 24 hours (until 9pm European, 3pm ET, 12pm PT Wednesday). Ask me anything.
I’m an ML/AI educator/founder. I got invited to the World Economic Forum in Davos. There's lots of politicians/investor-types but also some of the greatest scientists, researchers and builders (Andrew Ng/Yann LeCun among them) – AMA
byu/WillSen inFuturology
33 Comments
Hi Will, I agree the AI hype is getting ridiculous.
Has anyone spoken of open-source AI? It’s almost as powerful as the leading investor-funded models. I’d be nervous about pouring 10s of billions more dollars into the likes of OpenAI, knowing they haven’t got anything exclusive to sell, you can’t get for free elsewhere.
What are the best comments/statements you’ve heard so far from the experts you mentioned like Andrew Ng?
Which blockchain(s) and directed acyclic graph(s) are being discussed?
Any tips for a high school junior looking to pursue a career in research in a somewhat related field (cybernetics/brain computer interfaces, but with aspects of AI)? What should I look for in a college, who if anyone should I reach out to, what kind of things should I look to major in, ect?
Is the majority of excitement from the Exec’s towards the bottom line or top line for AI capabilities?
do you think there is AI now that the elite doesn’t want us to know about (or going to be)? and why don’t you stay for more than 24 hours? thanks for doing this ama
What’s the best way to kill this stupid wasteful AI fad before our oligarchs force its mass adoption to get a return on their investments?
Assuming widespread automation, what does the future economy look like to you? Are there any discussions around long term plans for what people will do in such an economy – UBI, public ownership of manufacturing etc?
Is there any palpable tension you can see between people and thinkers like Andrew Ng, who comes across like a pretty good guy with a realistic take on the dangers of AI used improperly, and the Open AI team, who seem to flood every event like this these days to tell everyone to give in without reservations to AI – in particular, their AI tools?
Any doc there?
When organ regeneration tech coming ? A tech like MatterGen that is released by microsoft past week‽
Any sus Oppenheimer there‽ >!use emojis in answer to say “yes” !<
I can see the “theme” is collaboration in the intelligent age, so what are notoriously uncollaborative companies like OpenAI saying in this vein? Or is that theme just something people ignore when they’re there and it’s more a sales event in reality?
i recently read an article (MIT) about how AI cant be used in health care (cancer detection) due to low amount of digitalization in healthcare, if we are going to make AI useful for humanity it should start with healthcare but it looks like we havent even digitized the industry at the first place. how are we going to use AI in healthcare more effectively in the future ?
Do you think AI will really change finance/accounting in the next five years beside better OCR?
Just popping in to say I think anyone who associates with these groups is evil.
It’s like watching Bond villains get together to talk about how much they hate their own species.
Hopefully AI is the solution rather than the gatekeeper.
Did you notice any non-experts (people who came in in other roles than “an expert”) who were actually surprisingly knowledgeable about AI?
Do you think that humans know enough about what we are to replicate the physics-based ability to appear to think? I mean this question in the ethical sense, and akin to a metaphor where “a human sees a dog, classifies it as a toothed mammal which relies heavily on its smell, and builds the physics pathways for teeth to be made, for a mammal to be made, and for an organism which relies on smell to be made. The humans engineer the physics-based paths which allow for matter to turn into these complex organized specifics. However, what comes out from these pathways are not a dog, but something else which fits the criteria, and is beyond our wildest horrors.” To simplify: we are creating the framework for a circle to be drawn by stating that a shape must be made which is made of all curved lines, and exists within the confines of this square. However, what comes out is a jungled mess of curves going inwards and outwards, not resembling a circle but still meeting the guidelines of our parameters because there are other things which can be made from the guidelines we give although we have no knowledge of such a thing even existing. We are creating what we think is human-based thinking mimicry, but we may be creating something entirely different which mimics this.
Is this level of thought and these type of questions common amongst the Davos crowd, and are these sort of philosophical and metaphysical questions what humans there are considering?
You say AI is dominating conversations at the expense of “other transformations” happening right now, what do you mean by that?
How do you walk around without wanting to slap every venal billionaire / hubris infused tech douche, in that epicentre of tax avoidance , the very control room of human misery throughout the world?
> WEF
>AI
I cannot express my honest thoughts about this combo without getting banned, but i do hope your jor gets automated and have the luck to live in a hypercapitalism world
Due to AI’s competitive influence on international industries, will data management and storage (that LLMs utilise)become more relevant to larger companies, potentially leading to information access and distribution issues?
Is there a way to make AI useful? So far it just gives me inaccurate information. I have to be able to recognize and avoid it so as not to lose my job.
Are there any discussions about the impact of AI IP on geopolitical power? Or any discussion about what to do with the companies reducing their costs, like increasing taxes to support those marginalized or something like that?
Among all the hyperbole and noise is there a graph emerging of the “new jobs” or areas that those taking recent and ongoing developments seriously can use to help focus skills/talent/efforts to actually emerge into the coming workforce landscape?
Too many people are all stood around slack-jawed wondering which direction to walk/run in…
The same Will Sentance who runs Codesmith? Why did you advise your students to lie about their work experience on their resumes? My company has stopped interviewing candidates with non-traditional backgrounds largely because of our experience with codesmith graduates deceptiveness. It’s a real shame for the rest of us self taught programmers.
Ai will write and review law proposals.
Its just text.
What IF AI looks in all of our law books and finds logical errors or simplification proposals.
Will governments let the AI review our law books ???
Is there any attention given to the level that guard rails degrade performance by making AI dependent on the expert configurators?
Most companies are currently ‘flirting’ with doing mass layoffs and replacing their staff with AI. Is mass unemployment being actively discussed as a bad thing to prevent, or a desired outcome as it will mean cost cutting for shareholders? Is it even desirable to replace your staff with a product owned by a different company that’s largely outside of your control? Thanks for doing this!
I think there’s been a slight an over-rotation from “AI will take all of our jobs” to “AI can’t do anything better than a human,” and that over the next few years (or less), we will really see how these tools start to drive layoffs – and not just in coding or copywriting. How transparent are the discussions on cutting headcount by adopting AI?
I work in tech and there’s definitely hype cycle at play (agreed re: “Agentic”), but also, AI can and will change the scope of many human-based jobs. Particularly areas like customer service feel vulnerable to this. We (or, ideally, all those people you’re surrounded by at Davos) need to plan for this immediately, IMO.
Is the obvious social dimension of the impending large scale roll-out of AI even a topic discussed in those circles?
LLMs are more than statistical models, yet are not able to reason, e.g., like a physicist. Is there a viable aproach on the horizon to overcome this barier?
It would be very interesting to know how much people on these kinds of forums think about the long-term consequences of automation and the introduction of artificial intelligence on the economy.
I assume that they are mainly discussing short-term and medium-term consequences of something like “Hey, look, we can automate so many jobs with this model and reduce costs by so many dollars!”, but they are not raising questions about how the market economy will work in the conditions of, say, mass automation of mental labor.
Yes, this is still a long way off, but the trends towards it are already visible.
Let’s think, Will, about what’s going on. It would be interesting to build small models and create a system that teaches people how to make those models to fit their needs, including model training. But that’s not a Davos issue. Everything there is geared toward keeping profits from getting away from the bottom line.
The security issue isn’t interesting either. I think the genie has already left the bottle.
The applications aren’t interesting either – despite the bravado, companies aren’t utilizing even 10 percent of today’s AI capabilities.
But ethical and business points aside, the question about model accuracy would be interesting. I’ve always wondered how local extremums are handled. Apparently, very poorly. For example, why not use the Monte Carlo method?
WEF is a snake pit of statist authoritarians. No idea why anyone gives a damn about their opinions.