Q&A
Staying Ahead of the Curve in Our New ChatGPT World
ChatGPT is here and it's not going to go away any time soon. Learn what you can do to stay ahead of the curve.
You can't escape it. The generative AI era is here. But how long will it last? And what importance does it have for enterprise IT?
Emile Cabot, eight-time Microsoft MVP and datacenter guru has jumped in head first into OpenAI's ChatGPT and has some valuable insight on how enterprises can benefit, even in the early days of the tool. He's ready to share what he's learned during his upcoming TechMentor conference session (taking place July 17-21 in Redmond, Wash.) during his session titled "Using ChatGPT as an IT Pro: Cool Tips and Infrastructure Prompts," a session that will be one that you won't want to miss.
Ahead of his talk, Cabot sat down with Redmond to break down why ChatGPT is more than just the next flash-in-the-pan tech trend and how you can start to use the tool today to make your job easier.
Redmond: So, the hype around "metaverse" came and went, just as it did for earlier buzzy tech like "blockchain," "Web3," etc. Not to say that these things are now obsolete -- far from it --but they're definitely in the "trough" part of the hype cycle. What do you think makes generative AI, and ChatGPT in particular, different from these?
Cabot: The main difference is the versatility of ChatGPT. Whether you're 5 or 95, an executive or a student, there is a benefit to using ChatGPT in many aspects of work and life. A closer comparison would be the search engine. The first time I sat in front of a search engine and realized I had a world of information at my fingertips, it took a few moments to figure out what to look for first. ChatGPT is the first tool that yielded that same experience. A wealth of knowledge, a wealth of possibilities. What do you do first? Ultimately for me, it was to create a poem for a family member. The second? A PowerShell script, of course!
The last couple of months have been wall-to-wall news around generative AI. Out of all the headlines -- from Microsoft's Copilot explosion, to Bing AI, to ChatGPT 4 hitting G.A. -- Â what do you think IT pros have to care about the most for the impact it will have in the months and years to come?
It's here, it's available to everyone, and it's not going away. In its present iteration, the benefit to our jobs is immense. Soon it will become fully integrated into our productivity suite and collaboration solutions, truly enabling easy access to generative AI in the workplace.
"It's here, it's available to everyone, and it's not going away. In its present iteration, the benefit to our jobs is immense. "
Emile Cabot, MVP System Center Cloud and Datacenter Management
Right now, at this minute, what's the best use case for ChatGPT in organizations? What is the thing that IT can start using it for immediately and see good results right off the bat?
Virtually every facet of an organization can see value in integrating ChatGPT. Composing e-mails, cleaning up code, brainstorming ideas on virtually any topic and performing data analysis. Through successful prompt engineering, ChatGPT can provide the opinion of a scholar, author, politician or religious figure. It can converse with a surgeon and offer alternative studies that align with a particular diagnosis, compiling data from multiple sources and presenting it in a conversational tone as if written by a person with a similar level of education.
What about security? Does ChatGPT pose a security headache for IT, and how do they overcome that? Conversely, how can they use ChatGPT to make their organizations' security posture even better?
Everything introduced to an environment poses a security headache for IT.
Currently, the big problem is that the main player in the game, OpenAI, can/does retain the prompts we input for use in further research and development, and there is no method to remove any confidential information that has been disclosed, despite GDPR regulations. This has become a huge headache for a microchip manufacturer that saw one of its engineers upload code to ChatGPT that contained proprietary information about its development and testing procedures.
Similar to Wikipedia, I expect organizations to begin deploying their own server infrastructure to host generative AI internally. While it generally will not be as comprehensive as ChatGPT in the public, the data sources, prompts and results will remain exclusive to the organization.
How important are prompts and prompt engineering? How can IT pros get better around prompts despite the technology around them changing so much almost every day?
This is, by far, the most critical skillset to develop when using ChatGPT. Subsequently, this is also the main reason why we do not fear that ChatGPT will replace our jobs. Although the tool processes natural language, the specifics of a prompt will vastly determine the usability of the provided response. Take the following example: If I ask ChatGPT to "create me a PowerShell script that installs the Hyper-V role, creates a VM, and then promotes that VM to a Domain Controller," it will provide me a script that's about 60 percent usable. However, if I rephrase the prompt to: "create me a PowerShell script that installs the Hyper-V role on a machine running Server 2019, creates a VM that is Internet-accessible, then uses PowerShell Remoting and the ActiveDirectory PowerShell module to install ADDS and promote the VM to a domain controller. For any steps of the script that require unique information, please provide variables at the top of the script for each item." I end up with something that's almost perfect.
Top ways IT pros can brush up their skills and increase their value in the ChatGPT era?
- Come to this conference! Some of the brightest minds in the industry from around the world speak at events like TechMentor, and the design promotes interaction between speakers and attendees.
- LearnPrompting.org has tons of mini courses on prompt engineering, prompt hacking, prompt tuning and more.
- Discord. The ChatGPT Discord community is huge. You can find a large repository of prompts, conversations about innovative use cases, cool ChatGPT hacks and more! (Check out DAN 2.0 for a fun favorite of mine.)
- Ask ChatGPT. No joke. Give it a prompt, then ask it how you could have better worded that. What additional information could have been provided to make the response more accurate? Ask it to give you a 20-minute hands-on exercise about prompt tailoring. The tool we're trying to learn has a vast database of information and the ability to present that data in a readable format. It already knows what we want to learn. What better teacher could you ask for at the wonderful price of free?
Job security is always a big concern, but especially now that AI has gotten more and more sophisticated and advanced. What does IT need to do to AI-proof their jobs? Any words of advice (or comfort) for IT pros worried that they'll be made redundant by AI?
And it will continue to get more sophisticated and more advanced every day. The key, like any emerging tech, is to stay on top of it. Unless your job is suited for the proverbial monkey, ChatGPT will be another tool in our arsenal. Fail to learn it, and, while you may not be in jeopardy of losing your position to generative AI, you will have to compete with someone that has learned to use it to augment their tasks.