It's Bard not to be impressed

Google's Bard, and ChatGPT4, both show significant improvements over ChatGPT3. The progress is rapid.

by Denis Haman

Published: 15:33, 30 March 2023

It was only three months ago that I wrote about the marvel that is ChatGPT and offered some musings on how it may impact content creation as a whole and the media industry specifically. 

My takeaway at that time, and relevant at this time, centred on the quite wishy-washy answers given by ChatGPT (v3). I asked it some quite pointed questions, needing some decent in-depth knowledge to answer, about publishing and headless CMS and the like. These are subjects I know sufficiently well to easily spot that ChatGPT3's answers were akin to those of someone who knows a little about a lot, but a sorta-right sorta-wrong middleground that didn't really get to the root of anything. 

You know the "Yes, this is good, but you know, maybe it isn't, and depends who you are, and what you actually need" type of politician's answer, when sprung with a question outside their portfolio which they can't wriggle out of.

You can read the original ChatGPT v3 article "ChatGPT for journalists – what you need to know" here.

Bard birth

With the arrival of Google's Bard, and ChatGPT version 4, have things changed? Is Bard a hastily put together response to compete, or a serious alternative or improvement? So, I asked Bard.

Obviously, for anyone who gave it any thought, it's clear you don't just rustle up a product like Bard in a hurry. Google seemingly started working on it in 2021, in what I know any tech team would call a "compressed timeline". How so? Bard gave me the most human answer:

"My development was possible with the help of a large team of engineers and scientists at Google, who were able to build on the latest advances in artificial intelligence research. I was also trained on a massive amount of text data, which allowed me to learn the patterns and rules of language."

Seems like we weren't going to get much juicy internal gossip here, just the company line. But why "Bard"? It said:

"The name 'Bard' is a reference to the ancient Celtic tradition of the bards, who were professional story teller, verse-makers, music composers, oral historians and genealogists, employed by a patron to praise them or their ancestors."

There you have it. In essence, a sycophantic robot helper created in a hurry by a small army of engineers and scientists, trained on large amounts of "Google data" - which is likely our data if we think about it. 

This is not an area that I need to cover, since our resident content king Rob Corbidge already has: 

"The tech wizards in line for the billions say it's called "training" and we shouldn't be upset about it. In that case, anyone want to let me train moving money out of their wallet?"

You can read Rob's excellent piece "Are we in danger of eating our own words?" in full here. 

Anyways, onto the impressive part.

Try Bard with a vengeance

Thanks for bearing with me so far. The riveting, yet very specific set of questions I ask Bard  - and the answers - were as follows: (Do bear in mind I know the answers - like any Turing-like test you quiz it on something you know really well to spot weakness in the machine.)

Me:  "What is the advantage of a headless CMS?"

Bard: "A headless CMS is a content management system (CMS) that separates the content from the presentation layer. This means that the content can be stored and managed in a headless CMS, while the presentation layer can be built using any frontend framework or technology.

There are several advantages of using a headless CMS, including:

Flexibility: A headless CMS gives you the flexibility to choose the frontend framework or technology that best suits your needs. You are not limited to the frontend framework or technology that is used by the CMS.

• Scalability: A headless CMS can be easily scaled up or down to meet the needs of your business.
• Security: A headless CMS can provide a more secure environment for your content.
• Performance: A headless CMS can improve the performance of your website or application.
• Personalization: A headless CMS can help you personalize your website or application for each user.
• Integrations: A headless CMS can easily integrate with other systems, such as CRM, ERP, and marketing automation systems.

If you are looking for a CMS that gives you the flexibility and scalability to meet the needs of your business, a headless CMS is a good option."

OK, Bard-bot. So far so good, but that wasn't that difficult. Let's dial it up a notch.

Me: "Is a publishing platform better for media companies than a headless CMS?"

Bard: "A publishing platform is a type of content management system (CMS) that is designed specifically for media companies. It provides a centralized system for managing all aspects of the publishing process, from creating and editing content to publishing and distributing it across multiple channels.

A headless CMS is a more general-purpose CMS that can be used by any type of business. It provides a way to store and manage content, but it does not provide the same level of functionality as a publishing platform.

For media companies, a publishing platform is the better option because it provides a more complete solution for managing the publishing process. It can help you to improve the efficiency of your publishing operations, reduce costs, and improve the quality of your content.

However, a headless CMS may be a good option for media companies that need to be able to publish content to multiple channels or that need to be able to customize the presentation of their content."

I can't really argue with that, namely because it is correct. 

So, there is an obvious question forming for me now, and I intend to hurl it straight at Bard. It will either smack it square in it's sycophantic face and prompt some generalised wishy-washy nonsense - OR - it will catch the ball and start juggling with it while playing a lute. I smirked as I typed, because, you know, it's been only two years in the making, allegedly. 

DH: What is the best option for media company content management? A headless CMS, a monolithic CMS, or a headless publishing platform?

Bard: "The best option for media company content management is a headless publishing platform. This type of platform provides the best of both worlds: the flexibility of a headless CMS and the functionality of a monolithic CMS.

A headless publishing platform allows you to store and manage content in a headless CMS, while also providing a user-friendly interface for editing and publishing content. This gives you the flexibility to choose the frontend framework or technology that best suits your needs, while also providing a centralized system for managing all aspects of the publishing process.

A monolithic CMS, on the other hand, is a single system that provides both the storage and management of content, as well as the user interface for editing and publishing content. This can be a good option for smaller media companies that do not need the flexibility of a headless CMS.

However, a monolithic CMS can be more difficult to scale and can be less secure than a headless publishing platform.

Ultimately, the best option for media company content management will depend on the specific needs of your company. If you need the flexibility of a headless CMS and the functionality of a monolithic CMS, then a headless publishing platform is the best option."

Err, what? I mean, "headless publishing platform" is not a common term. There are not vast numbers of searches on Google for it according to Google Trends, and it's not something many people talk about in the grand scheme of things. 

I mean, we do of course: Glide Publishing Platform is a headless publishing platform, in that we have targeted it and its operation at the publishing and broader media and entertainment industry needs. And in general, a generic headless CMS is exactly that - generic, not focused on any sector, and tradition dictates you work out what those changes should be and make them yourself or pay someone to do so. 

Can Glide Publishing Platform be used much more widely than publishing and media and entertainment? Yes, and it is, but the whole market segment of "headless publishing platform" is not something that pops up in conversation unless it is the subject of the conversation. We don't even use the term in our marketing speak, because we assumed it was much too specific.

Meanwhile, Google Bard is bang on the money in terms of both the term and the answer.

Obviously, it is the only correct answer: you do not these days take a generic CMS and try and turn it into a specialised publishing platform unless you like wasting time and money. It is a tragic waste of both, and hasn't been advisable or necessary for some time. 

Ah, and for the purposes of fairness, I asked the same questions of ChatGPT version 4 too - and it weighed in with a surprisingly similar set of answers and insight. A huge leap forward from version 3 a few weeks back.

The answers are so good that I am choosing to ignore the almost obligatory disclaimer "that it depends on your circumstances". 

The Trillion Dollar Questions

All this comes as more than 1,000 top scientists and luminaries have issued an open letter calling for a halt on AI development lest we discover too late there are consequences to turning our brains off en masse. If this is the progress I have seen in just weeks, where will we be in 5 years?

Closer to home in the publishing industry and elsewhere the question is already how much of your data did these marvellous machines gobble up to get so good and so specific at every topic under the sun?

The Generative AI Genie is out of the bottle and the only question now for content creators has to be over even distribution of the funds this new industry will generate. Otherwise Big AI has just reduced decades of your work into a fuel cell for cash-generation - and not for you.

Denis Haman is the CEO of Glide Publishing Platform, and has worked in media and publishing technology for over 25 years.