Chatbot Nirvana is About to come for ChatGPT

Breaking: Google Invests in AnthropicAI and Claude with $300 Million Round for 10 Percent of the A.I. Lab valued at $5 Billion

If you enjoy articles about A.I. at the intersection of breaking news join AiSupremacy?here. I cannot continue to write without community support. (follow the link below). For the price of a cup of coffee, Join 140 other paying subscribers.

https://aisupremacy.substack.com/subscribe

What it means for the future of Generative A.I.

Best time to Start an A.I. Newsletter

If you want to write a Newsletter in datascience, technology or A.I, I think you should just do it! In 2023, Substack is the best discovery, growth and retention platform to gain new audiences and build valuable niche communities.

I’ll be thrilled if you started yet another A.I. Newsletter, honestly the more points of view the better.?We’re all learning together.

Start a Substack

A.I and Software Development Jobs

You can visit the?archives here, the?web view here?and the?about page here.

• For?A.I. related jobs?go here.?Also join my?LinkedIn Group?where I hand-pick jobs and companies to help you find software engineering or machine learning opportunities.

Are we near Conversational A.I. Nirvana?

Hello Everyone,

While Microsoft has solidified its confidence in OpenAI and ChatGPT and presumably already has access to GPT-4 merging with its software, Google and others are not far behind.

According to?new reporting?from the Financial Times, Google has invested $300 million in one of the most buzzy OpenAI rivals,?Anthropic, whose recently-debuted generative AI model Claude is considered competitive with ChatGPT. Claude might actually be more advanced in terms of A.I. alignment and trust & safety.

Anthropic A.I. is a startup with an advanced Constitutional AI feedback interface, that extends what?RLHF is capable?of. A good primer for what reinforcement learning with human feedback is can be found?here. It’s basically what allows ChatGPT like products to be improved with human feedback.

Upgrade to Premium

Claude was in a closed beta in January, 2023 and had been seeking additional funding. This $300 million round from Google is a very big deal. Claude was created using a technique Anthropic developed called “constitutional AI.” As the company explains in a recent Twitter thread, “constitutional AI” aims to provide a “principle-based” approach to aligning AI systems with human intentions, letting AI similar to ChatGPT respond to questions using a simple set of principles as a guide.

Many believe that RLHF is the bottleneck that allows Conversational A.I. to get better, more trustworthy and more human.

Read the paper

There has been a lot of speculation if Google is falling behind Microsoft due to the speed of OpenAI and ChatGPT’s supposed success reaching 1 million users in five days and now?100 million users in just 2 months. This as Microsoft is likely to roll out ChatGPT in Bing any day now, with Google not many weeks behind with Sparrow or its LaMDA product.

LaMDA, which stands for Language Model for Dialogue Applications, is a family of conversational neural language models developed by Google and have been in development for quite some time. One of these is called?Apprentice Bard.?This is under the Atlas project. Now that Google has also invested $300 million in Anthropic A.I. for 10% of the company, this likely means the following:

• Microsoft?has the A.I. lab?OpenAI
• Google?will work more closely with?AnthropicAI
• Stability A.I.?will work more closely with AWS?and?Amazon.

Facebook’s Meta AI unit is also likely to release A.I. products as it has a long history of cloning and copying others. Meta is broadly facing increased competition from TikTok and a weakened targeting advertising system due to Apple’s 2021 iOS privacy update. It’s also facing the likes of Amazon, Microsoft and Apple beginning to encroaching on its duopoly with Google in digital advertising.

• Meta A.I. is likely to create A.I. products like ChatGPT
• ByteDance (the company behind TikTok) is also highly likely to create a ChatGPT competitor.

So when we think of Microsoft’s $14 billion investment in OpenAI and ChatGPT, and Google’s incredible response with as many as 20 new A.I. products coming in 2023 and 2024, we have to consider how Conversational A.I. will likely become ubiquitous with several serious competitors to ChatGPT. Microsoft is starting to?roll out ChatGPT into its other products.

Anthropic’s Constitution A.I. is the closest we have to laws of LLMs, reminiscent in an offhand way to Asimov’s laws of Robotics:

Source:?Wikipedia

Isaac Asimov was of course an incredibly influential science fiction writer. Asimov was considered one of the "Big Three" science fiction writers, along with Robert A. Heinlein and Arthur C. Clarke.

We are at a point where ChatGPT’s capabilities have been recognized. But the stark lack of fact verification and tendency to hallucinate can be fairly problematic, among other things. First movers like OpenAI rarely do win out in a new category if Conversational A.I. is indeed a new category. ChatGPT Plus is now charging $20 and in 2023 I think we are going to see at least four to six serious contenders emerge just in the coming months alone.

• OpenAI’s?ChatGPT
• Google DeepMind’s?Sparrow
• Google’s?Apprentice Bard
• AnthropicAI’s?Claude
• Baidu’s?Ernie. (ERNIE stands for?Enhanced Language RepresentatioN with Informative Entities)
• And around?4 others. (I would anticipate from Meta A.I. and ByteDance, as well as the other A.I. labs)

All the excitement we have felt since November 30th, 2022 since ChatGPT came out, we will get to feel again and test countless new Conversational A.I. and Generative A.I. tools and interfaces.

Founded in 2021, Anthropic also conducts research into AI language models but is also a spin-off of OpenAI in its own right. That AnthropicAI has thus raised $1 Billion in just two years is also astounding.

The Generative A.I. Battle over Search

The biggest anticipated battle between Microsoft and Google is anticipated to be between Bing and Google, with a Conversational A.I. layer of Search. Both Baidu and Microsoft are likely to launch these within mere weeks.

Google is ramping up its AI development and could have over 20 AI projects ready to be unveiled this year, according to a report by the NYT. Google is a far more applied A.I. R&D company than Microsoft, even if Microsoft Research has closed the gap in research years in pure research.

While Microsoft surprised many with the speed and size of its deal with OpenAI and ChatGPT’s popularity, it’s really just the beginning for the Generative A.I. phase of what LLMs will be capable of.

Google already has plenty of in-house expertise developing AI language systems, and?FT?suggests one motivation for the deal is simply building up Google’s cloud computing business. But it’s not even that, it’s because AnthropicAI is really thinking very hard how to make Conversational A.I. even more safe, truthful and aligned to human feedback.

Google wants to get things right, if it’s going to integrate it into its powerful and dominant Search monopoly and product. AnthropicAI have been doing important work to make Chatbots like Claude even more trustworthy and precise than ChatGPT:

AnthropicAI are using methods for training a harmless AI assistant through self-improvement, without any human labels identifying harmful outputs. If Google owns According to the reporting, Google will take a stake of around 10% at a $5 Billion valuation, this means Google has likely first access to some of the developments in A.I. alignment that Claude will likely be known for.

AnthropicAI?and not OpenAI might be a more pivotal A.I. lab in the long-run, due to their special emphasis on Building Reliable, Interpretable, and Steerable AI Systems (RIS AI).

AnthropicAI is a small, highly collaborative group of researchers, engineers, policy experts and operational leaders and their focus is more in the original spirit of OpenAI, not the commercialization that we have seen with Microsoft’s significant $10 Billion investment in them.

The organization was created in 2021 as a?public benefit corporation?by Dario Amodei, a former vice president of research at OpenAI. Claude itself might become a rival to ChatGPT in?ways we cannot yet anticipate in 2023.

I agree with the stark difference of Microsoft and Google in this regard. Many AI researchers have criticized the actions of OpenAI as increasingly reckless, particularly its launch of ChatGPT onto the public web late last year?without proper safeguards?or software that could reliably detect its output. We need to be cautious with Conversational A.I. as powerful as ChatGPT, not push the borders of safety on an already corruptive and Ad-centric internet where hype is often mistaken for truth.

Ironically it’s the real A.I. company, Google, and not Microsoft, that is the more cautious one, as it owns a share of around?84.08 percent of Search.?I don’t anticipate ChatGPT in Bing or Ernie in Baidu is going to make much difference. The?New Bing?could launch any day or week now. But a rush of ChatGPT competitors is coming and Microsoft is not certain to gain any substantial advantage in all of this inspite of its boldness in buying out OpenAI.

Pledge your Support

The competition to ChatGPT and Microsoft's so-called adoption of A.I. is going to get immense and it won't wait for 2024.

If you enjoy articles about A.I. at the intersection of breaking news join AiSupremacy?here. I cannot continue to write without community support. (follow the link below). For the price of a cup of coffee, Join 140 other paying subscribers.

https://aisupremacy.substack.com/subscribe