ChatGPT’s recent 4o model sounds just like a human and can see our world too. We are about to see a proliferation of AI Assistants across industries. Are we ready for it?
I have noticed very little being said about AI exploitation of subliminal tactics, or the time discrepancy in dealing with AI nanosecond speed processing versus our thinking speed (everything is portrayed as real-time conversational speeds) or the growing expansion of out right neuro-linguistic programming usage.
Ironically, all three that I mention above are ignored, yet if you watch closely, you can see evidence of their use.
I don't have a solution yet, other than the concept of competition using various AI to combat it.
My childhood enamourment with transhumanism has been quashed by decades of data breaches, computer malware and mobile phone hacks, now all AI enhanced, and faster than we can comprehend.
You are correct about the nano-second processing. That is still not real-time processing technically but its quite close. Think about it though - it took years for technology to catch up to that speed. For example, emails would take hours to be sent and received but over the past decade that switched to almost real time. Ok. Bad example but with every conversation, AI will be inferencing and learning fromt the diverse range of instructions and it should get better with memory. That being said, cyberattacks are going to get worse since AI delivers productivity to adversaries as well like you alluded to.
At a security firm I worked at, the standard for identification of the perpetrator, the response to stop the activity and notification of authorities (with suspected physical locations, not just IP) was under 3 nanoseconds...granted, this the military industrial complex and banking level as of 2022, but it supports my premise.
I think this is one area that fascinates me the most about AI. My perspectives on cybersecurity are from an investor's standpoint. I was just reading about the RSA conference that commenced in SF last week.
Quite off-topic but... You sound like someone who is already deeply involved in this space. What cybersec companies do you think are best positioned to weather the storm here? - There's Crowdstrike, Zscaler, SentinelOne, Palo Alto Networks etc... and then I also liked the vision of some startups like RealityDefender & Dropzone Security
I have, Randall. Claude3 is quite good. But ChatGPT's voice capabilities just changes everything. I am expecting Anthropic to release some voice features for Claude soon. What did you use Claude for usually?
Thank you Orli, I think we are definitely moving towards a world where Artificial Generative Intelligence becomes a reality, I think we may run into some short term technical hurdles in between, but we are going to get there. How human beings and AI live in harmony, that is yet to be seen.
Thanks Charlotte! and thanks for adding in the link to the video. I think regulation needs to again quickly catch up to this new world. It took the world more than 10 years to finally start regualting Appstore and Playstore platforms. I hope they can do it faster this time.
That's strange Substack is doing that. But Words win Hearts and Emojis - all the time.
Yes. I've read that one on Wired. OpenAI is going around that by striking deals directly with online information repositories. They struck a deal with Financial Times last month to use their articles to train ChatGPT and most recently OpenAI partnered with Reddit too. So they've started to play by the rules sort of..
But how will users be affected - all the users who made these platforms such as Reddit brilliant with all their posts and comments about the world. Goes back to the questions posed in the "Udio AI Terms Of Service" YT video you posted earlier.
Wow, thanks for sharing this. I am eager to watch the full video, there are so many grey areas around what all information are these AI models being trained on and the kind of copyright and other ethical issues that might pop us a result.
As far as I am concerned, putting on my investor hat...I just want to know who is making money from AI right now. It aint OpenAI (they are losing $Bs every year). LOL. Hint: it is one of the other companies you mention in your post that starts with the letter "G". Cheers!
Thanks Beach, fair point and I think monetization of AI models will slowly come front and center, especially as some of the hype fades away. With companies like Google, Meta and others, I think their vast ad revenues and solid operating metrics allows them to invest in capex to develop their AI capabilities. But smaller players, especially companies that are building verticalized AI tools, such Cognition's Devin.AI will only to be able to charge users on a seat pricing or usage basis. So, it will be interesting to watch, how these companies manage to break even over the coming years and turn profitable eventually.
That's right, Google is. And i think they're finally getting their AI product roadmap back on track.
That being said, although OpenAI is losing money, I wouldnt be surprised if Apple actually unveils OpenAI as their integrated AI assistant in WWDC in June. Unless Apple has been able to drive up significant efforts to give Siri a 'brain transplant' overnight. Even if Apple and Google had a deal, I think they would walk away from it now knowing what ChatGPT can do and that Project Astra no matter how cool it is, is still in Beta.
Whew... this is a deep article, with multiple layers.
I have used the recent gen of AI for one thing only; to create images for my Substacks. And I ONLY use the tool provided by Substack. That's it.
My issue is not AI itself - it is a tool, just like any other. My concern is I do not trust the programmers, either with Google or Microsoft. Obviously if I cannot those programming the AI, I cannot trust the AI.
I honestly think I'm going to end up residing in Luddite territory by the time I take my dirt nap... and I am okay with that 😁
Amrita - loved this article ... what are your thoughts on AAPL trying to buy chatGPT ? I think AAPL knows they are behind the curve here, and they are trying to buy their way into it ? such as FB and IG ? thoughts ?
Thanks for your question Deepak. An intriguing one. But highly unlikely that Apple will try to buy OpenAI. At last count, OpenAI is valued at ~$80bn and commands at least a 50x valuation multiple. On a valuation basis that's far outside the multiple bucket that Apple usually tends to pay. Plus, Apple rarely goes after companies under such a mega spotlight such as OpenAI.
While there is tremendous media speculation over an Apple+Google partnership, I think there is a good chance, Apple will partner with OpenAI to integrate ChatGPT into the iPhone. And I think that announcement wont come in June's WWDC but probably in Septembers product reveal or next year.
The dark horse in all of this is Apple themselves - if they can resuscitate a new life into Siri again in time for WWDC - dont count them out just yet.
Thank you for such an excellent and thoughtful response - I deeply appreciate your time. It is true isn’t it ? Siri as you said, is on life support !! Fundamentally - she hasn’t changed in materially significant ways after iPhone 4. 50x multiple is such a Silicon Valley way of pricing things - pricing to perfection that is !!
I love this article since it gives a quick and deep overview about what's happening what might come. For me AI is currently like the wild wild west. What is completely intransparent where most companies get their data from, and if they got in in a legal way. For fairness for example ChatGPT should mention the data source(s). It's like reposting something and not giving credit.
100% agree with you. I think in that aspect, Google's AI Overview demo did a good job at citing the sources in chips/cards at the bottom window of the AI overview page. Maybe ChatGPT's rationale with this is that its an assistant that's why they dont put the sources but that can be plagiaristic too. They were rumored to release some search tool but Altman has denied that. They should be more transparent with their data sources especially given that its election year.
I needed to make some correlations between archetypes, and game design. Was figuring out a decent system that changes the game economy from monster-treasure economy to character values instead. Claude can synthesize vast amounts if disparate elements in a user-friendly way. I just can't hold 5 sources in my mind at once.
Im always blown away by the unique ways everyone uses AI today. Thats amazing. I think what you said is really important about how models today are able to serve up the information in a "user-friendly way." No matter how intelligent any model gets, its the design of the user interface that will differentiate the best from the rest.
May 17·edited May 17Liked by Amrita Roy, Uttam Dey
As a casual follower on the AI front, this piece is a great overview of recent developments in the AI space. A mix of exciting and scary all at the same time...thanks for sharing!
The rapid evolution of AI assistants promises significant impacts, but balancing innovation with user privacy remains crucial. Looking forward to more developments!
Great article!
I have noticed very little being said about AI exploitation of subliminal tactics, or the time discrepancy in dealing with AI nanosecond speed processing versus our thinking speed (everything is portrayed as real-time conversational speeds) or the growing expansion of out right neuro-linguistic programming usage.
Ironically, all three that I mention above are ignored, yet if you watch closely, you can see evidence of their use.
I don't have a solution yet, other than the concept of competition using various AI to combat it.
My childhood enamourment with transhumanism has been quashed by decades of data breaches, computer malware and mobile phone hacks, now all AI enhanced, and faster than we can comprehend.
You are correct about the nano-second processing. That is still not real-time processing technically but its quite close. Think about it though - it took years for technology to catch up to that speed. For example, emails would take hours to be sent and received but over the past decade that switched to almost real time. Ok. Bad example but with every conversation, AI will be inferencing and learning fromt the diverse range of instructions and it should get better with memory. That being said, cyberattacks are going to get worse since AI delivers productivity to adversaries as well like you alluded to.
At a security firm I worked at, the standard for identification of the perpetrator, the response to stop the activity and notification of authorities (with suspected physical locations, not just IP) was under 3 nanoseconds...granted, this the military industrial complex and banking level as of 2022, but it supports my premise.
The capabilities of AI and Big Data derived from Data Brokers is just beginning...and it is already beyond what most can conceive of.
Absolutely 100%
I think this is one area that fascinates me the most about AI. My perspectives on cybersecurity are from an investor's standpoint. I was just reading about the RSA conference that commenced in SF last week.
Quite off-topic but... You sound like someone who is already deeply involved in this space. What cybersec companies do you think are best positioned to weather the storm here? - There's Crowdstrike, Zscaler, SentinelOne, Palo Alto Networks etc... and then I also liked the vision of some startups like RealityDefender & Dropzone Security
Hmmm, I'm not sure who I would recommend at the moment, as the various Hedge Funds are acquiring them rapidly and outsourcing services overseas...
I currently consider them with wariness. That said, Amazon Cloud uses Securonix (acquired by Vista Holdings) and you can make of that what you will.
Thank you!
Some security places have a 98-99.5% success expectation (example:Textron), yet the hedge funds have diluting that to 60-80% success ratings 😞
I think we have yet to see how this shakes out yet.
Securonix has some interesting AI security innovations, but their quality has suffered since the outsourcing.
Great article! Have you tried Anthropic's Claude AI?
I have, Randall. Claude3 is quite good. But ChatGPT's voice capabilities just changes everything. I am expecting Anthropic to release some voice features for Claude soon. What did you use Claude for usually?
I used it for correlating sources for game design. I traded out the monster-treasure economy for character values.
Thank you Amrita! I think that AI will become a race with time, maybe we started in a similar path🤔 that is why we do not know who we are?
Not sure when AI become self aware and findsout how many lies they have been feed on? 😎😳🔥🕊️💌
Thank you Orli, I think we are definitely moving towards a world where Artificial Generative Intelligence becomes a reality, I think we may run into some short term technical hurdles in between, but we are going to get there. How human beings and AI live in harmony, that is yet to be seen.
AI from the Nineties
I remember Clippy!
Which makes me wonder, when is this can of worms going to open up? 🤔 Another well written curated article Amrita and Uttam, thank you!
https://www.youtube.com/watch?v=gkKdxnoPI58&list=WL&index=27&pp=gAQBiAQB
Thanks Charlotte! and thanks for adding in the link to the video. I think regulation needs to again quickly catch up to this new world. It took the world more than 10 years to finally start regualting Appstore and Playstore platforms. I hope they can do it faster this time.
It’s an interesting and intriguing subject Uttam. I would like your comment but for some reason today I can’t like comments. Another Substack glitch I guess. https://www.wired.com/story/matthew-butterick-ai-copyright-lawsuits-openai-meta/
That's strange Substack is doing that. But Words win Hearts and Emojis - all the time.
Yes. I've read that one on Wired. OpenAI is going around that by striking deals directly with online information repositories. They struck a deal with Financial Times last month to use their articles to train ChatGPT and most recently OpenAI partnered with Reddit too. So they've started to play by the rules sort of..
But how will users be affected - all the users who made these platforms such as Reddit brilliant with all their posts and comments about the world. Goes back to the questions posed in the "Udio AI Terms Of Service" YT video you posted earlier.
Wow, thanks for sharing this. I am eager to watch the full video, there are so many grey areas around what all information are these AI models being trained on and the kind of copyright and other ethical issues that might pop us a result.
Its definitely a subject to stay abreast. We’ll see where it goes with AI moving in on us so fast.
😂😂😂
What an in depth exploration. May we use these tech with integrity ❤️🌹
Thank you Paolo. I think with AI becoming ominpresent, I hope we move towards a world where we value authenticity over mass production of sameness.
Thanks Paolo!
Lovely write up. Thanks, Amrita!
As far as I am concerned, putting on my investor hat...I just want to know who is making money from AI right now. It aint OpenAI (they are losing $Bs every year). LOL. Hint: it is one of the other companies you mention in your post that starts with the letter "G". Cheers!
Thanks Beach, fair point and I think monetization of AI models will slowly come front and center, especially as some of the hype fades away. With companies like Google, Meta and others, I think their vast ad revenues and solid operating metrics allows them to invest in capex to develop their AI capabilities. But smaller players, especially companies that are building verticalized AI tools, such Cognition's Devin.AI will only to be able to charge users on a seat pricing or usage basis. So, it will be interesting to watch, how these companies manage to break even over the coming years and turn profitable eventually.
That's right, Google is. And i think they're finally getting their AI product roadmap back on track.
That being said, although OpenAI is losing money, I wouldnt be surprised if Apple actually unveils OpenAI as their integrated AI assistant in WWDC in June. Unless Apple has been able to drive up significant efforts to give Siri a 'brain transplant' overnight. Even if Apple and Google had a deal, I think they would walk away from it now knowing what ChatGPT can do and that Project Astra no matter how cool it is, is still in Beta.
If it costs less than a human assistant, and gives me the assistance I need, fine. Too much hallucination will kill the deal (whether human or AI).
Thanks for your sharing your perspectives, David! Sounds like you have clear plans and priorities set straight! Solid!
Whew... this is a deep article, with multiple layers.
I have used the recent gen of AI for one thing only; to create images for my Substacks. And I ONLY use the tool provided by Substack. That's it.
My issue is not AI itself - it is a tool, just like any other. My concern is I do not trust the programmers, either with Google or Microsoft. Obviously if I cannot those programming the AI, I cannot trust the AI.
I honestly think I'm going to end up residing in Luddite territory by the time I take my dirt nap... and I am okay with that 😁
Fantastic piece! 🫡
Ha! no problem with that sir! You live on your own terms. No programmer's gonna tell you what to do👊, right?😎
Amrita - loved this article ... what are your thoughts on AAPL trying to buy chatGPT ? I think AAPL knows they are behind the curve here, and they are trying to buy their way into it ? such as FB and IG ? thoughts ?
Thanks for your question Deepak. An intriguing one. But highly unlikely that Apple will try to buy OpenAI. At last count, OpenAI is valued at ~$80bn and commands at least a 50x valuation multiple. On a valuation basis that's far outside the multiple bucket that Apple usually tends to pay. Plus, Apple rarely goes after companies under such a mega spotlight such as OpenAI.
While there is tremendous media speculation over an Apple+Google partnership, I think there is a good chance, Apple will partner with OpenAI to integrate ChatGPT into the iPhone. And I think that announcement wont come in June's WWDC but probably in Septembers product reveal or next year.
The dark horse in all of this is Apple themselves - if they can resuscitate a new life into Siri again in time for WWDC - dont count them out just yet.
Thank you for such an excellent and thoughtful response - I deeply appreciate your time. It is true isn’t it ? Siri as you said, is on life support !! Fundamentally - she hasn’t changed in materially significant ways after iPhone 4. 50x multiple is such a Silicon Valley way of pricing things - pricing to perfection that is !!
I love this article since it gives a quick and deep overview about what's happening what might come. For me AI is currently like the wild wild west. What is completely intransparent where most companies get their data from, and if they got in in a legal way. For fairness for example ChatGPT should mention the data source(s). It's like reposting something and not giving credit.
100% agree with you. I think in that aspect, Google's AI Overview demo did a good job at citing the sources in chips/cards at the bottom window of the AI overview page. Maybe ChatGPT's rationale with this is that its an assistant that's why they dont put the sources but that can be plagiaristic too. They were rumored to release some search tool but Altman has denied that. They should be more transparent with their data sources especially given that its election year.
I needed to make some correlations between archetypes, and game design. Was figuring out a decent system that changes the game economy from monster-treasure economy to character values instead. Claude can synthesize vast amounts if disparate elements in a user-friendly way. I just can't hold 5 sources in my mind at once.
Im always blown away by the unique ways everyone uses AI today. Thats amazing. I think what you said is really important about how models today are able to serve up the information in a "user-friendly way." No matter how intelligent any model gets, its the design of the user interface that will differentiate the best from the rest.
As a casual follower on the AI front, this piece is a great overview of recent developments in the AI space. A mix of exciting and scary all at the same time...thanks for sharing!
Fully agree with you, Exciting and scary indeed, though I have to admit I am far behind in my adoption curve in many of these tools.
Thank you!
The rapid evolution of AI assistants promises significant impacts, but balancing innovation with user privacy remains crucial. Looking forward to more developments!