The other issue is every company is trying to market their "Generative AI" story to their customers in the recent months -- even when the company does not really have any worthwhile AI technology which is useful. Somehow every marketing department believes they need to push their "AI story" -- which has many customers laughing in disbelief when the actually try the immature or non-operational "Generative AI" technology out pushed in the pitches. Taking a look at a closer example -- there are multiple products being pushed recently as plug-ins to ChatGPT which claim they can generate Powerpoint presentations for you saving you hours of labor. The office workers who regularly have to generate Powerpoints love the concept but the reality falls far short. They quickly find out that most the the generated Powerpoints are next to useless and the only way to generate a useful Powerpoint would require hundreds of hours of work to tailor the input prompts. Of course this amount of work is far more many hours than simply the eight to sixteen hours that a worker would normally need to create a Powerpoint presentation manually.
But will the chatbot be a succes? Not meaning to be irritating but incorporating a chatbot is not the same as succesfull employment. I personally havent seen or used a really succesfull one.
There is another thread in Science and Technology covering the recent events involving the CEO shakeup at Open AI in the posts since Friday. https://www.elitetrader.com/et/thre...ure-of-information.372136/page-8#post-5894117
Chat Bots tend to just piss off customers -- who spend their time demanding to be connected to a real customer support rep -- or the equivalent if on the phone.
Yeah, and what then? I also believe that the essence of AI is overrated, because nowadays there are no obvious reasons to state that it's somehow helpful and ease our daily life. Well, of course, some people use ChatGPT, for instance, others admire how robots work on the Amazon warehouse, but in fact, toda ytis' too early to claim about revolutionary impact from AI in all spheres. AI development has two outcomes, and I guess you understand what I mean, either it will be for the sake of the society, or...
AI is a catch phrase like Cloud or Crypto. It doesn't mean anything other than describing a process that already existed and is just being furthered by better hardware (nVidia GPUs, high core count CPUs via ARM, Terabit networking, PCIe SSDs, etc.). Facebook, Google, Microsoft, Amazon, etc. were already working or using bits of it already. Then there is Elon who was already deploying a decision tree analysis from the nVidia GPUs for his FSD. GM & Google have been testing their self driving taxis here in the SF for a while now (although GM's license has since been revoked). It does feel to me like the Netscape IPO moment in terms of software & hardware ramp up. It will benefit the industrial sector in the end the most I believe via robotic automation. I mean Elon is already doing some of that at Tesla. I worked in Pharma IT for about 19 years and all the software and databases we created sped up the drug discovery cycle especially finding pathways to search for similar organic reactions (Phase I). Pharma workers were pretty software savvy but the Biologics workers weren't so we never were able to crack that market. That's probably the next frontier for automation Biologists & MDs. Doctors are notorious Luddites.
Useless? Slidebean seems to differ. Says it can get to 90% of same output as humans. Granted, it looks like he may be selling something but the question is where the final output rates from useless to 90%. I'm guessing closer to 90%. But be interesting to see an objective 3rd party evaluation.
As a consumer I cannot see the day when AI directly makes my life better. Maybe companies I deal with will make fewer errors when I order stuff, maybe they will be more responsive when I report problems or glitches, but I doubt this is life-changing nor significantly more economical. Maybe AI might have a bigger impact for me personally in public services. Fewer civil servants would be a great step forward.
Everyone seems to have a friend doing amazing things with AI lol. Such obvious bullshit. I still use chatGPT4 everyday but the magic has quite worn off. It is so easy to see how incredibly limited it is after much use. If anything, what is amazing about AI at this point is how incredibly useless all of this has been outside of NVDA selling compute. Then as these models scale up and get better, we will reach a point that the cost of compute is simply uneconomic and it will still just make sense to pay the cheaper human to do the work instead. It is the same mistake we made thinking how robots would take all these blue collar jobs 30 years back. Like today you could have a warehouse full of boston dynamics robots at 20X the price of the human warehouse workers and give your competitors a massive edge in labor cost.