Passer au playerPasser au contenu principal
  • il y a 22 heures
VTNews D2 1400 AI eats software

Catégorie

🤖
Technologie
Transcription
00:01Hello and welcome to VivaTech News, I'm Marjorie Payon
00:04and you're watching Breaking Tech, your daily dose of innovation
00:07here live from VivaTech.
00:09It's definitely 2pm here in Paris and probably somewhere else
00:13coming up at the top of this hour.
00:15Startup valuations continue to plummet in 2023.
00:19Well, there's room for hope.
00:21We'll navigate the VC seed stage world in a few.
00:25AI is bubbling here at VivaTech.
00:27We'll get an update with our favorite human expert.
00:31And actually this favorite human expert is of course sitting with me
00:35here on this very set, Rudy Rutenberg.
00:39I'm definitely going to need some of your knowledge about AI on day 2.
00:43Yesterday we were just trying to understand how President Macron
00:47especially would actually embrace the AI revolution in terms of regulation too.
00:52but overnight actually the EU Parliament met further steps when it comes to the AI Act.
01:00This is definitely the piece of European regulation everyone has on its lips here at VivaTech.
01:06But it's not expected until the end of the year because due to the European political process
01:12now we need to go to Trilog.
01:16Europe obviously has a decision making process that starts with the Commission then the Parliament then the countries have their
01:21say.
01:22What was interesting of course is that yesterday we had the input from President Emmanuel Macron
01:27and he was indicating that he was not so much in favor of an implemented regulation.
01:33he was arguing for something more like what was done around the international fiscal process which happened within the OECD.
01:40That is obviously a much longer process than something that would be put in place quite quickly, relatively quickly I
01:48would say, by the European Union.
01:51The discussion here is very mixed.
01:53Some people are saying yes we do need regulation, some people are saying okay well let's first focus on the
01:59immediate issues with AI.
02:02For example avoiding bias, making sure that AI gives the right answers first of all because we know that AI
02:08right now
02:09because you know it's trained to provide you, we're talking about the larger language models here for example they are
02:15trained to provide you with an answer.
02:17If they don't have an answer right now they are just making stuff up so we need to fix that.
02:22And at the same time for example there was Siemens, the Siemens CTO today on stage who was saying let's
02:29be very careful with regulation of AI.
02:32At the same time you have Google, the head of the AI business who is saying well actually we welcome
02:37it because then we know what we need to design.
02:40And he was saying we already put in principles in 2018, we're talking about things like do no harm.
02:47Because yeah, AI can go to many places.
02:50Yeah but when I think of Sam Altman now, OpenAI CEO pushing for more regulation and more actually pieces of
02:59legislation to regulate its own creature,
03:02then I'm just actually saying like you Rudy or thinking like you Rudy that yeah probably it's better for this
03:08type of scale ups and big tech to know what they have the right to do or not and play
03:14by the rules to create further development when it comes to AI.
03:18Speaking of which, is generative AI already in the past?
03:22That's basically what Matt as Yann LeCun was saying yesterday on stage one.
03:28Well what is different with generative AI is that it affects knowledge jobs, right?
03:37Because we've had the industrial revolution where we saw that manual labor was affected.
03:42Now for the first time we have a technology that basically has an impact on mental jobs.
03:50Mental jobs sounds a little weird but...
03:52It is.
03:53But I'm not going to, you know what?
03:55Knowledge jobs.
03:55You're human so still.
03:57Knowledge jobs.
03:58And that is something that we haven't dealt with.
04:02And again there, is it going to free up people to be more creative?
04:06Or is it going to hurt jobs?
04:10When the Siemens CTO was asked this direct question, he sort of avoided the issue.
04:16And he says it's going to basically stop people from having to do boring repetitive tasks.
04:22But he didn't really say what it's going to mean for jobs.
04:25So now, what the matter with Kansas?
04:27I mean what the matter with actually fostering AI in Europe?
04:31We do have the knowledge, we do have the scientists, but we don't have the labs.
04:36Right now what is interesting is indeed that we are lacking the computational facilities.
04:42So the computing power to train these large language models.
04:45And large language models are one of the very exciting areas of AI right now since chat GPT-3.
04:54And all of these models are coming from the United States.
04:58So we have obviously OpenAI, you mentioned them.
05:02We have Google.
05:03We have IBM.
05:04We have Microsoft all working on these.
05:06But there's nothing in Europe.
05:08We have now for the first time a French company, Michel AI, who just founded a month ago
05:16and raised 105 million euros already.
05:19They want to be the developer in France of a generative AI company.
05:26So we'll see where that goes.
05:28Absolutely.
05:29And we'll continue actually following pretty closely.
05:31AI Stories here at VivaTech with you.
05:34Rudy, thank you so much for your know-how.
05:37And now, Total Energies is committed to expanding its renewable energy portfolio,
05:42especially in the offshore wind energy sector.
05:46To do so, the French multinational oil and gas company has partnered with Nash,
05:51a Dutch offshore wind developer.
05:53A Dutch offshore wind developer.
05:55Thank you.
Commentaires

Recommandations