Why Are OpenAI and SoftBank Investing in “Human Understanding”?

When I saw the partnership between OpenAI and SoftBank, I found myself thinking about something slightly different from the usual conversation around AI.

Most discussions focus on infrastructure, business strategy, data, or geopolitical competition.

But I could not stop thinking about something else:

How far is AI trying to go into human understanding itself?

Not simply language processing.

Not translation accuracy.

Not productivity.

Something deeper.

For a long time, AI development focused on information.

How accurately can language be processed?

How naturally can responses be generated?

How efficiently can knowledge be organized?

But recently, the direction seems to be shifting.

AI is beginning to move toward areas that are far less structured:

• hesitation

• silence

• emotional ambiguity

• desire

• loneliness

• traces left inside language itself

In other words:

the unstable parts of being human.

This is why I do not think the relationship between OpenAI and Japan is only about technology or market expansion.

Japanese communication often demands something beyond literal correctness.

Not just what was said.

But:

• how smoothly it was said

• whether hesitation existed

• whether the feeling feels lived through

• whether understanding seems to have “passed through the body”

That last part may sound strange.

But I think it matters.

In Japanese, there is a word: shikkuri.

It roughly means “to fit” or “to settle naturally,” but the sound itself seems to carry part of the meaning.

The slightly heavy rhythm of “-kkuri” feels as if something catches slightly in the throat before settling into place.

The word itself carries resistance and settling at the same time.

And perhaps understanding works similarly.

Sometimes, understanding does not arrive as explanation.

It runs through the body first.

The phrase may sound slightly unusual in English.
But that discomfort is part of the point.

Understanding is not always clean or perfectly organized.

Sometimes it arrives more like a current, a shock, or a small internal movement before language fully catches up.

This may also explain why Japanese communication can feel difficult to fully translate.

Not because Japanese is “mysterious,” but because meaning is often distributed across things that are not entirely verbal:

• pauses

• rhythm

• indirectness

• timing

• atmosphere

• physical intuition

Even silence can carry meaning.

Sometimes, not speaking immediately is itself part of communication.

This is where I think the conversation becomes larger than language.

The real question may not be:

But rather:

That question feels increasingly important in the age of AI translation.

Interestingly, I think modern culture itself has already begun moving toward a different kind of understanding.

Music is one example.

With DAWs (digital audio workstations), music creation changed dramatically.

Music no longer always moves in clear linear progressions.

Instead, sounds are layered, revisited, edited, fragmented, rearranged.

Emotion no longer arrives only through structured buildup.

Sometimes it appears suddenly, partially, non-linearly.

People often describe modern media consumption as fragmented attention.

But perhaps something else is happening too:

understanding itself may be becoming more non-linear.

We no longer always understand things from beginning to end in a perfectly sequential way.

Instead, we move back and forth:

• short clips

• fragments

• commentary

• atmosphere

• recommendations

• AI summaries

• scattered impressions

And somehow, meaning still forms.

Not as a straight line.

But as an edited constellation.

This may also explain why “smoothness” alone no longer guarantees trust.

In fact, perfectly optimized language can sometimes feel strangely empty.

Too complete.

Too frictionless.

Human understanding often leaves traces behind:

• pauses

• unevenness

• searching

• revision

• emotional residue

Perhaps this is why people still look for signs that understanding was genuinely experienced rather than merely produced.

That may also be why the OpenAI × SoftBank partnership feels symbolically important.

Not simply because of technology.

But because AI may now be approaching the point where language alone is no longer enough.

And if AI eventually attempts to understand humans more deeply, then Japan may become an unexpectedly important place to observe.

Not because Japan is superior.

But because Japanese communication often forces the question:

For a long time, modern systems were built around what might be called continuous understanding:

• linear learning

• structured explanations

• sequential logic

• stable interpretation

But perhaps humans were never entirely like that.

Perhaps we only prioritized that model because it was easier to organize, teach, and measure.

Now, something else is becoming visible again:

• interrupted understanding

• emotional jumps

• bodily intuition

• fragmented meaning

• non-linear connection

Not as errors.

But as part of understanding itself.

As multilingual translation spreads across platforms like note, we are no longer being asked only about translation accuracy.

We are being asked something much stranger.

What remains after translation?

What survives when language changes?

And how much must truly align before humans feel: