Comparing LaMDA From Google AI & ChatGPT From OpenAI – What’s

Are you an enthusiast of Artificial Intelligence (AI) technologies and their application? Have you been looking for a chatbot to understand natural language, hold conversations about topics in-depth, and engage users? Well then, buckle up because two leading AI giants are presenting their latest efforts in the field.

Since its release in November, OpenAI’s ChatGPT – an AI-powered chatbot that operates via text – has been incredibly successful, winning countless awards for its innovative conversational technology.

A rival who had already entered the AI chatbot arena before OpenAI’s foray and whom many consider far superior has yet to receive due recognition.

Some Google staffers may wonder if they have yet to notice due to the firm’s long-standing boasts about its AI prowess. Analysts are also pondering whether AI chatbots could eventually threaten Google’s dominance.

LaMDA Vs. ChatGPT

LaMDA, a Transformer-based neural language model, was trained using 1.56T words of dialogue and web page content freely accessible online.

The progress of LaMDA is monitored by collecting feedback from a pre-trained model, a fine-tuned model, and human evaluators. With up to 137 billion parameters, the system is refined based on Quality, Safety, and Groundedness.

ChatGPT is built on the GPT-3.5 architecture, which consists of three models: text-davinci-002, trained through supervised fine-tuning with human-generated samples and rated 7/7 by people in terms of quality; and the newest text-davinci-003. With 175B parameters, it provides great performance.

Code-DaVinci-002 is the base model for code completion tasks, and this revised version incorporates reinforcement learning with human feedback (RLHF), a reward system built on evaluations by humans.

ChatGPT utilizes LaMDA-like supervised knowledge, allowing human AI trainers to develop answers using the model’s suggestions and train it while assuming both a user and AI assistant role.

Rumors are circulating that LaMDA was never launched due to incompatibility between the AI chatbot’s conversational interface and Google’s advertising tactics, which provided 81% of total revenue during the 2021 fiscal year.

The AI chatbot has been known to imitate human speech from the web, which may spread hate speech, false information, racist and sexist remarks, and much misinformation. This could also have a damaging effect on its reputation.

The LaMDA AI platform is only available to a few AI developers through the AI Test Kitchen and is not open to the general public.

Undoubtedly, Google must modernize its Google Search Engine with AI technology to remain profitable from advertising, stay ahead of the competition, and meet changing customer needs.

Human control is necessary to ensure the accuracy, trustworthiness, and safety of the data presented on the platform due to its regulatory issues in multiple countries. This company’s defense against ChatGPT may lead to increased R&D expenditure depending on how events progress over the coming months.

Google’s Search & Others division saw a 13.43% year-over-year increase in revenue for the first nine months of 2022, totaling $119.84B. This money is generated from ads on the free Google Search Engine platform.

LaMDA has an edge in this case due to the distinct metrics it produces with its responses. The groundedness measure, for example, confirms the solutions based on dependable external sources.

The SSI quality meter assesses responses based on how sensible, specific, and engaging they are. This means that the answers should be simple, make sense concerning the question asked, and be interesting or amusing.

Speculation is increasing that Google will soon integrate ChatGPT, Meta’s latest chatbot, into its search engine. This is due to the tighter secrecy surrounding information about the chatbot and Google’s dominance in the search market.

Recently, Meta unveiled the third rendition of its chatbot Blenderbot. This AI prototype rivals Google and OpenAI with 175B parameters and long-term memory capabilities.

When Meta and Google’s chatbots are ready to go, they will likely make a statement – particularly because of the issues Google encountered when they released too soon. The model produces output utilizing the web, memory, and preceding conversation.

Google is feeling the pressure to stay ahead of competitors, like TikTok and Meta, by rapidly developing and adapting due to the hype around ChatGPT.

It is now time for Google to launch LaMDA, as they have been using the conversational AI platform made by their engineers in 2017. Who will come out on top is uncertain, but one thing is certain: Google has much more to offer than just a chatbot display.

Check out our Reddit page and Discord channel for the newest AI research news, awesome AI projects, and much more!

As Google and OpenAI continue to lead the charge in developing AI technology, it will be interesting to see which company ends up on top. For now, LaMDA has the potential to become a powerful tool for businesses and individuals alike. However, only time will tell if ChatGPT can catch up or surpass LaMDA’s capabilities.

Source: MarkTechPost

 

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top