A Brief History of LLM Chatbots

Hello friend! Are you fascinated by the rapid advancements in AI chatbots powered by large language models (LLMs)? As a fellow tech enthusiast, I wanted to provide an in-depth look at four of the top options – Bing Chat, ChatGPT, Chatsonic, and Google Bard.

There‘s so much hype and confusion around these conversational AI systems. My goal here is to give you the details you need to evaluate each chatbot clearly, based on their unique capabilities, limitations, and ideal use cases. Let‘s dive in!

First, some background. How did we get from simple Q&A systems to AI assistants able to debate philosophy?

The journey started around 2018, when research from groups like OpenAI showed initial promise in large language models. LLMs are AI systems trained on massive text data sets, allowing them to generate remarkably human-like text.

OpenAI‘s 2018 LLM called GPT already impressed with its 1.5 billion parameters. But the 2020 release of GPT-3 represented a true breakthrough. At 175 billion parameters trained on 570GB of text, GPT-3 could churn out essays, articles, stories, and even code based on simple text prompts.

Clearly, LLMs had potential for conversational AI. In response, Microsoft and other tech giants started developing specialized LLMs for dialog applications.

Microsoft built the LLM TuringNLG leveraging GPT-3 advancements, which later evolved into ChatGPT. After training TuringNLG on human conversation data, ChatGPT launched in 2020. But only in November 2022 did ChatGPT become open access.

Around the same time, Google also recognized the promise of LLMs for language interfaces. They developed LaMDA in 2021 – a 137 billion parameter LLM tailored for dialog. LaMDA now powers Google‘s new Bard chatbot.

Other tech firms and startups also jumped in, like Anthropic with its reputable Claude assistant, and Writesonic‘s development of Chatsonic.

This sets the stage for today‘s commercial battle between conversational LLM chatbots. Next let‘s do a technical comparison!

Here is an overview of how Bing Chat, ChatGPT, Chatsonic, and Google Bard stack up regarding key features:

	Bing Chat	ChatGPT
Platform	Microsoft Edge browser only	Web portal + iOS/Android apps
Free vs. Paid	Free	Free + $20/month ChatGPT Plus
Needed Registration	None	Required
Query Limits	30 per chat, 300 daily max	No hard limits
Underlying AI Model	Microsoft Proprietary LLM	GPT-3.5 and successors

	Chatsonic	Google Bard
Platform	Web portal + iOS app	Built into Google Search
Free vs. Paid	Free tier + paid plans	Free
Needed Registration	Required	Google account login
Query Limits	No hard limits	None yet
Underlying AI Model	GPT-3.5 + Sonantic	Google LaMDA LLM

Let‘s explore some key similarities and differences:

Platform and Availability

Bing Chat‘s integration directly into Microsoft Edge gives it seamless access for Windows users. However, Edge has only about 4% browser market share worldwide as of February 2023 according to Statcounter. So only a tiny slice of users can access it easily.

ChatGPT and Chatsonic support all major browsers via their web portals. But ChatGPT also offers iOS and Android mobile apps, allowing use on virtually any device or platform. This gives it the widest access.

Google Bard has only been implemented in Google Search so far. But since Google accounts for over 90% of search traffic, this gives Bard huge reach immediately.

Free vs. Paid Plans

Bing Chat and Google Bard are currently free for all users, with no paid tiers or extras.

ChatGPT remains free as well, but offers a ChatGPT Plus premium subscription for $20 per month. This provides priority access when servers are overloaded, faster response times, and early previews of new features.

Similarly, Chatsonic has a free account option, but advanced features require a paid plan starting around $20 per month. These include higher usage quotas, faster responses, and access from multiple devices.

Registration and Login Requirements

Bing Chat stands out for requiring zero registration – just start chatting in Microsoft Edge. The others all need some initial signup:

• ChatGPT: Requires an OpenAI account using email and password

• Chatsonic: Needs an email and password to create account

• Google Bard: Needs to be logged into an existing Google account

No registration makes Bing Chat the path of least resistance. But the others allow personalization, conversation memory, and user preferences linked to accounts.

Limits on Queries and Usage

Perhaps the biggest limitation of Bing Chat is the strict quota of only 30 queries per chat session, and a max of 300 queries per day. Hitting these limits constantly interrupts conversations.

ChatGPT and Google Bard currently have no hardcoded limits on daily queries or conversations. However, ChatGPT may throttle usage during server overload.

Chatsonic also has no specified limits, but does restrict abuse based on a vague "fair usage" policy.

So for free-flowing conversations, ChatGPT and Google Bard offer the most flexibility.

The Underlying LLM Technology

Each of the four chatbots relies on proprietary LLMs engineered by the parent company:

• Bing Chat: Microsoft‘s own proprietary LLM, likely based on GPT-3 foundations.

• ChatGPT: Originally leveraged GPT-3, but now runs on more advanced successors like GPT-3.5 developed exclusively by Anthropic.

• Chatsonic: Uses text generation model Writesonic, reportedly built off a fine-tuned GPT-3.5 model. Incorporates Sonantic‘s conversational engine as well.

• Google Bard: Leverages LaMDA, Google‘s massive 137 billion parameter conversational LLM. Likely combines other proprietary models too.

Most experts agree ChatGPT has the most capable underlying LLM technology today. But Microsoft, Google, and others are racing to improve their models.

Given everything we‘ve covered, let‘s summarize the unique pros, cons, and ideal use cases for each AI chatbot:

Bing Chat

Pros

• Tight integration into Windows through Microsoft Edge
• Completely free with no account required
• Powered by robust proprietary Microsoft LLM

Cons

• Very limited reach since Edge lacks market share
• Restrictive caps of 30 queries per chat and 300 per day
• No personalization or conversation history

Best for

• Windows users already on Edge seeking an AI chat intro
• Very casual testing of conversational AI with minimal setup

ChatGPT

Pros

• Leading accessibility via web and mobile apps
• No hardcoded limits on queries or chats
• Accounts allow personalization and preference tracking
• Advanced GPT-3.5 foundations handle complex conversations

Cons

• Free tier access can be throttled during peak demand
• Responses are polished but may sometimes be inaccurate
• No inherent safeguards against false information

Best for

• Testing capabilities of a cutting-edge LLM chatbot
• Creative writing and brainstorming applications
• Exploring trends in artificial intelligence

Chatsonic

Pros

• Generous free tier with few limits
• Combines GPT-3.5 text generation and Sonantic conversation models
• Paid subscriptions provide more capabilities

Cons

• Less name recognition than ChatGPT or Google
• Free tier lacks priority access and advanced features
• Smaller user base and community support

Best for

• Cost-conscious users wanting quality AI chat
• Unlocking higher production volumes with paid tiers
• Commercial applications via its API and integrations

Google Bard

Pros

• Integration into dominant Google Search engine
• Validates Google as an AI leader after lagging behind
• Leverages Google‘s own LaMDA LLM and resources

Cons

• Still in very limited preview mode
• Lacks standalone chat portal for conversations
• Heavily optimized for search assistance over chitchat

Best for

• Early adopters wanting to try Google‘s entry
• AI aid for search queries and information lookup
• Showcasing Google‘s progress with LLMs

With this comprehensive feature comparison, which AI assistant is the right match for you? Here are some quick tips:

• For Windows fans, Bing Chat is the path of least resistance. But beware its strict query limits.

• If you want the most advanced conversational model available, try ChatGPT. It strikes the best balance today.

• Prefer saving money but still want quality? Start with Chatsonic‘s generous free tier.

• If you primarily want AI help while browsing and searching online, test out Google Bard. But expect integration hiccups.

Of course there‘s no unambiguous "best" – consider your priorities around platform support, features, use cases, and budget. Each chatbot has distinct pros and cons suited for different needs.

Today‘s LLM chatbots are exciting, but still noticeably flawed. However, their capabilities are advancing at unbelievable pace.

Most experts agree we‘re still in the very early days of conversational AI. The big tech firms are investing billions into the next generations of LLM technology.

We can expect capacities to grow by orders of magnitude in the coming few years. And developers are also focused on enhancing critical areas like:

• Personalization – Maintaining individual user context and preferences.

• Multimodal skills – Incorporating visuals, audio, and video alongside text.

• Verifiability – Enabling fact-checking by citing sources rather than guessing.

• Specialization – Developing expertise in niche domains like medicine, law, etc.

LLM chatbots today provide a small glimpse of the future powered by natural language human-computer interaction. Still, buckle up, because rapid progress is coming!

Let‘s recap the key conclusions:

• Bing Chat, ChatGPT, Chatsonic, and Google Bard represent leading commercial LLM chatbot applications, each with distinct pros and cons.

• ChatGPT appears superior for general use based on its platform reach, advanced technology, and flexible free access.

• Bing Chat offers easy AI conversation access for Windows Edge users, albeit with very limiting quotas.

• Budget-minded folks can tap into Chatsonic‘s generous free tier before paying to upgrade.

• Google Bard focuses on search assistance over free-flowing chat, given its search integration.

• We‘re still in the very early days of LLM chatbots. Expect exponentially faster progress in coming years!

I hope this overview has provided ample context to evaluate these chatbots yourself. Let me know if you have any other questions! I‘m always happy to chat more about AI technology and speculate on the future.