What do I need to know to take advantage of AI?

​

Basically nothing. I will take care of it for you, and you will be able to use it like you use other software programs.

But, if there is something you would like to know, I can help you learn.

​

Should I learn to prompt to get responses from an AI model?

​

Effective prompting requires a good deal of education. However, you don't need to be an expert to get value out of AI. Basic prompting is easy - you can get the basics here. The basics may not help you implement AI processes in your business. But you will be amazed at how including AI in your workday enhances your effectiveness. Quickly, you will wonder how you did without ChatGPT or Claude or whichever GPT you choose (I use several, because they have different strengths and provide different answers). Always keep in mind that GPTs hallucinate, so you must double check any results that you will be relying on or publishing.

​

Is hallucination a big deal? What about the systems I will be purchasing from you?

​

It can be a big deal, particularly with regard to research and drafting papers. We will use techniques, including prompt construction, to minimize hallucinations, and LLMs are trying to minimize them. However, hallucinations are not so much a bug as a consequence of the ways LLMs are constructed. Hallucinations cannot be eliminated - at least not now. Having said that, I do not recall the last time I identified a hallucination in my work with AI, and I will provide you with ways to look for hallucinations.

​

What about data privacy and security?

​

The internet is inherently insecure. If you would like or require privacy and security beyond that found every day on the internet, including HIPPA and GDPR, we can implement solutions either on your premises or within a cloud server that will increase your data privacy and security.

​

I am told there is a danger that my data could be used for training a large language model and therefore become public. Is this a risk?

​

Yes, it is. However, your solutions will be implemented in such a way as to eliminate that risk.

​

Everything AI is changing so fast, it is difficult to keep up.

​

Absolutely. We are at the beginning of a revolution. A lot of money and effort are pouring into AI. It is important to ignore all the hype and the shiny new objects and focus on long-term value. While it is impossible to know who the ultimate winners will be, I will be using leaders in each category of solution and will only employ niche solutions after discussions with you. My newsletter and YouTube channel will help keep you up to date.

​

What about the Chinese products like DeepSeek and Manus?

​

The Chinese are producing impressive and innovative products that will strongly compete with American products (the French are also competing at a high level). These Chinese products come with significant data safety and privacy issues. In general, I will not use these products in our solutions. As long as you are aware of the data safety and privacy issues, these products are worthwhile exploring on a personal basis - I use them daily. If you would like to use them in a company solution, we can discuss alternatives such as self-hosting.

​

What do you think about AI taking over human jobs?

​

I am blown away by what AI can accomplish today, and this is just the beginning. I am in the camp that thinks that a significant number of jobs will be lost to AI over time. I also think that those who know how to have AI help them in their jobs will become much more effective in their jobs and much more competitive in the job marketplace. The history of the introduction of new technology is one of continual job destruction accompanied by the creation of even more new jobs enabled by the new technology. I hope that happens this time, but I wonder. AI is becoming part of our lives very quickly, and AI doesn't sleep at night or complain or take vacation. And, it can speak multiple languages, is able to find any bit of information at a moment's notice and can compile and assemble breathtaking amounts of data essentially instantly. During the internet and manufacturing-robotics era, those who lost jobs were told to "learn to code." Ironically, AI is now automating the generation of code at evermore sophisticated levels - an increasing number of coders will lose their jobs. Many people dance around this issue because they do not want to cause stress. I believe these issues should be understood without sugar--coating. I think the adjustments caused by AI will be significant and perhaps profound.

​

What kinds of problems will I encounter when implementing AI solutions?

​

Implementing AI will be much like implementing any software system - learning curves, practice, refinement, bug shooting. Since AI is generally intuitive to operate, this should not be much of a challenge. Many companies will encounter resistance and some fear on the part of their employees, often mixed with fascination and anticipation. It will be important to educate employees, to make them part of the process and to deal with any negative perceptions.

​

How can I tell what LLM is best for me?

​

The answer changes almost daily. You should pick one-three as your go-to LLMs and stick with them unless you have a reason to change.  One approach is to go to LMArena.ai, which continually updates scores for different LLMs for different tasks.

Having said that, unless you are getting deeply into a particular type of task, pick one of the majors - ChatGPT is always a safe choice - and don't worry about it. Since you should think about using one LLM to check or augment your primary LLM, you might choose a second. At this writing, I primarily use Perplexity, Grok and Manus, and I have Claude running locally on my computer, but that will certainly change.

​

Do you take care of me during set up and while the systems are running?

​

Yes, I implement the solutions with you. The reason for recurring charges is to provide maintenance to the products and to provide customer support.

​

Do I own the solutions you provide for me?

​

I provide you with a worldwide, exclusive, nontransferable license for your specific implementation. ​

​

Do you need access to my computers?

​

Yes, I will be accessing your computers remotely to develop and install your solutions, including interfacing them with your existing systems.

​

​What is MCP?

​

Model Context Protocol is the next big thing in enabling the use of agents and interacting with your applications. It particularly helps with on-premise solutions.

​

What is GEO?

​

Generative Engine Optimization, and its twin AIO, or Artificial Intelligence Optimization, are beginning to intrude on SEO, Search Engine Optimization, which makes websites more responsive to Google's rankings.

There is a trend towards people not going to websites, but using GPTs to answer questions by scraping the information off of websites. GEO is updating websites to better serve the so-called crawlers, bots that visit pages across the web to collect the data that informs the answers provided by AI.

The idea is not so much to be found and ranked by SEO as having information readily available to GPTs so that their answers to users well represent the information on your website.

We can discuss this in greater detail - I even have a bot that will take a look at your website and tell you how GEO you are.

​

How about A2A?

​

AI systems are evolving rapidly. Today, organizations are automating and AI-izing their companies. The next step is for one company's AI agent to be able to work with another company, letting AI agents talk, work, and share tasks across apps, clouds, and companies. The Agent2Agent protocol has been established to enable this next step.

We will get there, but we need to walk before we run.