AI With Omnilow

I believe Omnilow algorithmic language is the most effective language for people to interact with AI in many contexts, for two AIs to communicate in some cases, and for a single AI to use within itself in certain situations. In those use cases, Omnilow would be more efficient and effective than natural language, computer language, and/or standard algorithmic language. This post explains more.

Writing in Omnilow is much quicker and easier than writing in a computer language. Here is Omnilow's main description.

Omnilow algorithmic language would be the best possible language for humans to use to interact with AI in some use cases for both consumers and developers. Among many other examples, Omnilow would very often be the best possible language for a consumer trying to get an answer to a very complex question where natural language is too ambiguous, and Omnilow would often be best for a developer to use to try to have an AI correct its behavior.

Omnilow would be the best possible language for AI to AI communication in some use cases. If that communication would otherwise be in natural language generated by one of those AIs, I think Omnilow would almost always be the better language for such communication. However, both AIs would need to understand Omnilow for that to work. Although, I think computer language would usually be the better language for AI to AI communication if it's reasonable to use computer language for such communication.

Omnilow would be the best possible language for a singular AI to use within itself in some use cases. Such a use case might be saving data for long-term storage and using a smaller amount of storage. My above points about using Omnilow for AI to AI communication would also apply for a single AI using Omnilow internally.

Extra Info

LMs have demonstrated an ability to understand and write Omnilow with decent accuracy from just a single prompt. More thoughtful LM models seem to do better at that than their quicker counterparts since (as of writing this and to the best of my knowledge) no LM has been formally trained on Omnilow. If any good LM was formally trained on Omnilow, avoiding the quicker model wouldn't be necessary, and I'm very confident that it would actually understand and write Omnilow with more accuracy and less computation when compared to natural language regardless of how much of an expert it is at natural language. For an AI, processing natural language naturally requires more computation than an algorithmic language. I'm also confident that formally training an LM to understand and write Omnilow would be easy if it's already good at natural language.

Here's a post I made about how Omnilow would be much more efficient than natural language in backend pipelines for LMs: SS5.org/2025/06/30/NL-vs-omnilow-LMs.

Another example of a use case where Omnilow would often be best would be where a consumer is trying to get a written complex system (to be understood and/or implemented by humans) and is written in a language that is more ambiguous (and thus less tedious) than a computer language or traditional algorithmic language. My Algorithmic Proof That You Should Rush post acts as a specific example of that. (There wasn't any AI involved in that.) Without Omnilow, that algorithm would either be too ambiguous to understand or much much more tedious to write and read.

If you read and understand the Omnilow algorithm in that post and then compare what it would have been like to try to make that algorithm in natural language, I'm very confident that you would agree that Omnilow is the best possible language that that algorithmic model could have been written in for both the writer and the reader.

Even if AI became better than humans at inferring data from natural language, Omnilow would still often be superior to natural language. Inferred data from natural language is often going to be wrong regardless of how smart the AI might be. For many complex systems, almost nothing from the understanding can be wrong or it throws everything off.

For example, "There are two companies. One company has loans which is all financially sound. For the other company, it's not financially sound." has many different possible meanings. It could mean that (the first company and its loans) was all financially sound and the other company also has loans and (the other company and its loans) is not financially sound. It could mean that the first company wasn't necessarily financially sound but its loans are financially sound and the other company (which doesn't necessarily have any loans) is not financially sound. And, many other combinations of different possible meanings.

Video On That Example

It's the finished algorithm I wrote in the video above. I'd recommend watching the video first.

(
{This is a continued example from SS5.org/2025/02/06/AI-with-omnilow.

This Omnilow algorithm will be a conversion from "There are two companies. One company has loans which is all financially sound. For the other company, it's not financially sound." into Omnilow.}

((there is)((Possibility #1)(or)(Possibility #2)(or)(Other Possibilities)))

{The above algorithm is a complete "There are two companies. One company has loans which is all financially sound. For the other company, it's not financially sound." into Omnilow.}

((Company A)=(((some)(specific))(Company)))

((Company B)=(((some)(specific))(Company)))

((Company B)(is)((not)(Company A)))

{That is "There are two companies." converted into Omnilow.}

(
(Possibility #1)=

((Company A)(has)(((multiple)(Loan))(which is)(Financially Sound)))

{That is "One company has loans which is all financially sound." converted into Omnilow.}

((Company B)(has)(((multiple)(Loan))(which is)((not)(Financially Sound))))

{That is "For the other company, it's not financially sound." converted into Omnilow.}
)

(
(Possibility #2)=

(((Company A)(has)((multiple)(Loan)))(which is)(Financially Sound))

{That is "One company has loans which is all financially sound." converted into Omnilow.}

((Company B)(is)((not)(Financially Sound)))

{That is "For the other company, it's not financially sound." converted into Omnilow.}
)

((Other Possibilities)(is)((not)((Possibility #1)((and)(or)(or))(Possibility #2))))

((Other Possibilities)=(((some)(specific))(((large)(Integer))(of)(Possibilities))))

{There are many different possibilities that those sentences could mean based on the ambiguity of English and natural language.}
)

There are other ways natural language can be impractical. For instance, inputting a principle as being equal priority to another principle instead of a higher priority to that other principle might lead to a totally different output. As a specific example of that, a stock management system might have 50 if-then principles that are interconnected, and some might conflict with each other.

Once again, as a more specific example of that, the system might say to sell stock if that individual stock composes over over 10% of the portfolio's net value, but also have a principle to buy a stock if it's considered undervalued. It'd be easy to clarify which one takes priority for just 2 such principles, but that becomes extremely tedious for 50 interconnected principles if one is using natural language. The ambiguity compounds to result in both the writer and reader having a much more tedious experience or a much less accurate understanding.

For less complex systems, Omnilow might still be the best language. For instance, the USA Constitution was written in natural language (and is thus ambiguous). A large part of the Supreme Court's job is simply to try to guesstimate what the Founding Fathers meant when they wrote what they wrote. That wouldn't be a problem if the Founding Fathers used Omnilow. (Although, not having digital writing/reading platforms such as Notepad++ makes Omnilow much less useful which is why someone before me didn't invent Omnilow.)

SS5.org/Sam-Omnilow is my dedicated page promoting myself for championing and enabling Omnilow solutions for LM (Language Model) companies. If that's relevant to you, I'd appreciate it if you'd check it out. Here's my post with free Omnilow all-purpose and personal investment portfolio prompts to copy and paste into an LLM of your choice. I'd appreciate it if you tried out one of those prompts. Any of those prompts instruct ChatGPT how to write and read/understand Omnilow so you or another AI can send Omnilow inputs and/or get Omnilow outputs.

SS5.org/2025/02/06/AI-with-omnilow