Inside modern business, as dem don dey use generative AI, e don pass just say na tool dem dey use am, e don enter stage wey dem go arrange am well-well inside di system.
Beyond dis, new kind of intelligence go show face: "Symphonic Intelligence."
Dis article go look into how generative AI dey used now and wetin e fit be for future, from two sides: work wey dem dey repeat (iterative work) and work wey flow well (flow work).
Work Wey Dem Dey Do Am Again And Again (Iterative Work)
For one article wey we write before, we examine di way wey "work wey dem dey do again and again and di tools dem" take different from "work wey dey flow and di systems dem." We see dis as ways to make generative AI fit do business tasks.
Work wey dem dey do again and again na when human beings dey combine different small-small tasks without even knowing, and dem dey try am out, make dem see how e go be.
And for dis kind of work, tools dey perfect. If you pick di right tools for different tasks, di work go move fast and well. So, e important to get all di tools wey you need ready and sabi use dem well.
For now, when dem dey use generative AI for business, most times na as a tool dem dey use am.
Many of di gists about how generative AI go make business better usually mean say dem just dey add dis new, strong tool to di ones wey human beings don dey use for work wey dem dey do again and again.
Di Wahala Wey Dey With Iterative Work
But on di other side, as we point out for di previous article, di kain improvement wey tools dey give for iterative work no too plenty.
As tools dey become more efficient, na human beings go come be di problem. At di end, you no fit pass di limit of how many hours human being fit work.
Wetin still dey, na big difference for how veteran staff dem and new staff dem dey efficient and accurate for iterative work, and e hard to close dis gap. So, even if you wan double di amount of work by next month, you no go fit do am if you no get people wey get veteran skills.
To solve dis problem of human beings being di bottleneck, di final solution na to replace everything with artificial intelligence.
But, di generative AI wey we get now never reach dat level of performance.
Still yet, even di iterative tasks wey small for eye, if you look am well, na many tasks wey dem dey do without knowing say dem dey do am.
Because of dis, dem no fit turn am into normal IT systems or manuals wey anybody fit follow, so dem just dey depend on wetin human being sabi do.
Unless dem arrange all dis plenty unconscious tasks wey need skills, and turn di necessary know-how for each one into knowledge, generative AI, no matter how much e improve, no fit replace human work.
How To Change Am Into Flow Work And Systemization
To solve di problem of sharing tasks within di current power of generative AI, and to arrange tasks wey we dey do unconsciously and turn wetin we sabi into proper knowledge, e very important to arrange di "try and error" kind of work (iterative work) into standard "flow work."
Dis standard flow work no just fit tools, e fit systems too.
Inside flow work, some tasks na generative AI go do, and some na humans go do. If you connect dem with a system, di whole flow work go fit run.
When you change work to flow work and put am into a system, e go give you many good results:
One, generative AI go special for individual tasks, so e go clear how to make generative AI work beta and more accurate for each task.
Two, many workers go fit add wetin dem sabi to di generative AI, and everybody go gain from am.
Three, e go easy to slowly shift di share of tasks inside dis work to generative AI.
Like dis, by changing iterative work to flow work and saving di knowledge wey generative AI need for each task as a system, intellectual work go dey move close to automation, just like for factory line.
And as generative AI dey improve well-well over time, and dem dey use di knowledge wey dem don gather for different tasks, e go come possible to make di whole flow work an automated process, with generative AI in charge.
Virtual Intelligence
Dis one na di end of di analysis from di side of iterative work and tools, and flow work and systems.
Another article wey I just write dey explain dis discussion more.
For dat article, I talk small about orchestration by virtual intelligence.
For now, and for di nearest future, because of how generative AI dey perform, e dey work better for efficiency and accuracy when e focus on particular tasks.
So, as we discuss am before with flow work and systems, di best way to do am na to connect specialized generative AIs for each task through a system.
But, even if generative AI performance too improve, e fit be more efficient and accurate to process by changing roles and using different knowledge inside one processing run, instead of just processing different tasks at di same time.
Dis method go remove di need for a system to link generative AIs together. Operations wey resemble system integration go dey happen inside di generative AI itself.
Wetin still dey, from a situation where you no fit rearrange or add tasks without changing di system, di generative AI itself go fit respond anyhow e like.
Dis mean say di tasks wey don turn to flow work and don dey systematized go go back to iterative work.
But, di iterative work wey go come back after e don pass through dis flow-working and systematization process go dey for a state where dem don form knowledge wey dem fit use again, even if dem increase di number of generative AIs or change their versions.
Dis one go solve di wahala of human iterative work and go make am possible to do flexible tasks wey resemble di ones wey humans dey do.
Here, I dey call di ability of generative AI to switch roles and knowledge during one execution "virtual intelligence." Dis one just like a computer's virtual machine.
Just like virtual machine technology dey form different computers wey dey run on one hardware, one generative AI dey process by switching between many roles.
Generative AI don already naturally get dis virtual intelligence power. Na why generative AI fit do discussion wey get many people or create novels wey get many characters.
If dis virtual intelligence power improve and dem give am enough knowledge, e go fit do iterative work.
Intelligence Orchestration
Wetin still dey, I dey call di ability to freely combine many roles and knowledge to do tasks "intelligence orchestration."
Dis one just like orchestration technology wey dey handle many virtual machines.
Just as orchestration technology dey run systems well by starting di virtual machines wey dem need when e reach time, a generative AI wey get better intelligence orchestration skills—wey be part of virtual intelligence power—go fit do iterative work anyhow e like, still maintain how fast and correct e be, while e dey use plenty roles and knowledge well-well.
Symphonic Intelligence
Generative AI wey reach dis level, dem fit call am Symphonic Intelligence.
Just like orchestra, wey each musician sabi play im instrument well-well, dey play one song while dem dey do their different parts, Symphonic Intelligence fit play di symphony of intellectual work.
Dis Symphonic Intelligence na new idea, e show di final level for generative AI.
But, Symphonic Intelligence itself don already dey.
Na we human intelligence be dat.
Na because we get Symphonic Intelligence make we fit do complex intellectual tasks anyhow we like, without even knowing, through iterative work, by using plenty of wetin we sabi.
Lastly: How AGI Go Be
If dem give generative AI, wey fit copy Symphonic Intelligence, di processes for flow work and di knowledge base for other tasks, e go fit handle many different iterative tasks.
As e come fit handle plenty different iterative tasks, e go likely grab di common rules and how knowledge dey structured across those tasks.
By dat time, for iterative tasks wey e no even know before, with just small explanation from human, e go fit learn wetin e need to know for dat task just by watching how human dey do am.
Dis na di real Symphonic Intelligence. Once e reach dis level, humans no go need to dey stress demsef to arrange work to flow or to put wetin dem sabi into code again.
Wetin still dey, di knowledge wey generative AI go gather automatically like dis, dem fit share am among demsef.
When dat happen, generative AI go sabi learn pass human by far.
Dem fit say dis na one form of AGI.