When Will Autonomous Semi Trucks Reach Industrial Scale?

These autonomous big rigs were going to only be allowed in interstates up to the transitions.

In otherworldly, a real person ( legally licensed or not ) would be required to drive / navigate the truck to the "on-ranp [ on transition ]" then turn the autonomous vehicle lose to terrorize the unsuspecting driving public.

Then, wnen these things reach their destinations, once again the truck would only be allowed the take the "off-ramp [ exit transition ] to the end of the transition where once again a real person ( legally licensed or not ) would be required to drive / navigate the truck to the final destination.

Personally, I think these things need special markings / lighting so the driving public knows these terrifying devices are around us!

But, what would I know?

Current evidence from autonomous driving cars is that, while the software causes some problems, it saves others, for a considerable net benefit in safety. I don't expect that relationship to come out differently for big rigs.

Not...A...Fan

With the commercial autonomous vehicles (WAYMO, ZOOX, NURO …. ) we the people can usually see and identify them. Not so with Commercial trucks pulling trailers, especially when approaching or overtaking them from the rear. Hence my comments about identifying the vehicles as being “autonomous”.

And then there's my muddled meanderings.... that follow...

But the bigger problem is as the automation and AI become more prevalent, the financial stability of the United States

The important point is the problem is not simply “people lose jobs.” It is that the U.S. tax-and-benefit system is heavily built around human labor income. If AI and automation eventually produce substantially more economic output with substantially fewer workers, the current financing model could become strained even if the country becomes much wealthier.
The According to the Congressional Budget Office (CBO) specifically identifies this possibility: if AI permanently displaces workers or pushes them into lower-paying jobs, income-tax and payroll-tax receipts could fall, while spending on income-support programs could rise. At the same time, higher productivity, corporate profits, new industries, and higher wages for some workers could increase tax revenues. The ultimate balance is uncertain. According to the Congressional Budget Office

Social Security is probably the clearest vulnerability

Social Security is primarily financed by payroll taxes on workers and employers.

In 2025, payroll taxes brought in about$1.7 trillion, equal to roughly 5.8% of U.S. GDP. Social Security payroll taxes alone are projected at about 4.2% of GDP under current CBO projections. According to the Congressional Budget Office

Imagine an economy where:

100 people + machines produce what 150 people used to produce.

If 50 workers disappear from the labor force, there are fewer paychecks from which Social Security taxes are collected. But the retirees don't disappear. Their benefit obligations remain.

That creates a fundamental mismatch:

Current system: Human labor → wages → payroll taxes → Social Security benefits

Highly automated economy: AI/robots → production → corporate/owner profits → ? → Social Security benefits

The question becomes: How do you tax the second chain to replace the first?

This is especially significant because Social Security already faces a long-term financing challenge independent of AI. CBO projects that Social Security expenditures continue to grow relative to GDP while revenues remain comparatively stable. According to the Congressional Budget Office

AI could therefore arrive at the same time that the existing demographic problem is already putting pressure on the system.

Medicare faces a similar problem Medicare Part A is also substantially financed through payroll taxation.

So automation could create a double effect:
- fewer workers paying payroll taxes;
- potentially more people needing government-supported health

However, AI could also reduce health-care costs through better diagnostics, drug discovery, automation, prevention, and administrative efficiency. And if AI makes people healthier and allows them to work longer, that could increase tax revenues. So Medicare's outcome is particularly uncertain.

CBO explicitly notes that AI could simultaneously increase revenues and spending—for example, if improved health care allows people to live longer, producing more taxable income but also more years of Social Security and Medicare claims. According to the Congressional Budget Office

Unemployment and welfare spending could increase during the transition. This could actually be the first major fiscal problem. Suppose AI eliminates 10 million jobs, but the economy eventually creates 8 million new ones. The remaining 2 million people aren't just an economic problem; they potentially become a government expenditure problem.

The government could face greater demand for:
- unemployment insurance;
- Medicaid;
- Affordable Care Act subsidies;
- SNAP;
- housing assistance;
- disability programs;
- job-training programs;
- tax credits such as the EITC;
- other means-tested assistance.

CBO specifically identifies increased participation in federal income-support programs as a possible consequence of AI-driven employment disruption. According to the Congressional Budget Office

And the timing matters enormously. If displaced workers are unemployed for six months, the fiscal problem is manageable. If millions remain economically unnecessary for 10–20 years, the U.S. would be confronting a fundamentally different social structure.

Income taxes could become less reliable. The federal government receives enormous amounts of revenue from individual income taxes. If labor's share of national income falls substantially, that tax base changes.

Consider two hypothetical economies:
Economy A — human-intensive
$30 trillion of economic output:
- $20T wages
- $7T business profits
- $3T other income

Economy B — highly automated
$30 trillion of economic output:
- $10T wages
- $17T business/capital income
- $3T other income

GDP hasn't changed. But the distribution of taxable income has changed dramatically. That matters because wages and capital income are not taxed identically.

CBO emphasizes exactly this issue: AI could shift income among workers, businesses and investors, changing federal revenues even if total economic output rises. According to the Congressional Budget Office

Corporate taxes could partially replace lost labor taxes—but probably not automatically. This is one of the most important counterarguments. Automation doesn't make the money disappear. If an AI system replaces 1,000 employees and the company saves $80 million in wages, that money may become:
- higher corporate profits;
- dividends;
- capital gains;
- lower prices;
- higher investment;
- higher wages for remaining workers;
- new business formation.

Some of those things are taxable.

So a highly automated economy doesn't necessarily produce less government revenue. The problem is that the tax system would have to capture a sufficient portion of the new income.

CBO specifically notes that increased corporate profits, dividends and capital gains could produce additional revenue, potentially offsetting some decline in labor-related taxes. According to the Congressional Budget Office

State and local governments could be hit particularly hard. This aspect is sometimes overlooked!

States and municipalities depend heavily on:
- individual income taxes;
- payroll-related economic activity;
- sales taxes;
- property taxes;
- business taxes.

If a community loses large numbers of well-paying jobs, you can get a cascade:

job losses → lower household income → lower consumption → lower sales-tax revenue → declining property values → weaker property-tax growth → higher demand for public services.

A city whose economy depended on 100,000 relatively well-paid workers could have a serious fiscal problem if AI reduced employment to 60,000—even if the remaining companies were producing more output than before. And states cannot generally print money to cover the gap.

There is a potentially much bigger issue: who owns the machines? This may ultimately be more important than the technology itself. Imagine that AI produces an enormous increase in productivity.

There are two very different possibilities.
Scenario A: Broad ownership
--------------------------------------
Millions of Americans own shares in the companies, AI infrastructure, robots, pension funds and other capital producing the output. Then people can receive income from capital even if they work fewer hours.

The economy could transition toward something like:
less wage income + more capital income + shorter working hours. That could be relatively prosperous.
Scenario B: Highly concentrated ownership
---------------------------------------------------------
Suppose AI and robotics produce enormous wealth, but most of the productive capital is owned by a relatively small percentage of households and corporations.

Then you could get:
very high productivity + enormous profits + declining median wages + fewer jobs + increasing inequality.

GDP could look spectacular while the median household feels economically insecure. That is the scenario that poses the greatest challenge to traditional American social policy.

The U.S. may eventually have to tax something other than labor. This is where I think the long-term policy debate becomes particularly interesting.

If machines do an increasing proportion of economically productive work, the government has several broad choices.

Option 1: Tax labor less
------------
Reduce payroll taxes and perhaps income taxes on workers. Then obtain government revenue from other sources.

Option 2: Tax capital more
-----------------------------------
Increase taxes on:
- corporate profits;
- dividends;
- capital gains;
- large estates;
- investment income.

This would shift part of the tax burden from labor toward capital.

Option 3: Create a consumption/value-added tax
-----------------------------------------------------------------
The government could collect revenue from economic consumption rather than primarily from wages. This would be closer to the taxation systems used in many other advanced economies.

Option 4: Tax automated economic activity
---------------------------------------------------------
This is the idea commonly called a “robot tax” or automation tax. For example, a company replacing 500 employees with an automated system might pay a tax intended to compensate for lost payroll-tax revenue.

There are major problems with this approach, however. It could discourage productivity-enhancing investment and would be extremely difficult to define consistently.

Option 5: Create a social wealth fund
-------------------------------------------------
The government could acquire ownership stakes in productive assets and use the returns to finance citizens' benefits.

Instead of: worker → payroll tax → Social Security
you might have:
national wealth → investment returns → citizen dividend/social benefits.
Alaska's Permanent Fund is a small-scale example of the basic concept, although its purpose and structure are quite different from a national AI wealth fund.

Option 6: Universal Basic Income or something similar
-----------------------------------------------------------------------
If machines eventually produce enough economic output, the government could redistribute a portion of that output directly to citizens.

The economic logic would be:
machines produce wealth → government taxes some of that wealth → citizens receive income independent of employment. That would represent a very significant change from the traditional American social contract.

There's an even deeper possibility: the tax system could become disconnected from employment. This is the scenario worth thinking about.

The traditional U.S. social model implicitly assumes:
Adults work → earn wages → pay taxes → receive benefits when they become sick, unemployed or old.

But suppose AI eventually makes full-time employment unnecessary for a large portion of the population.

Then we could have:
Machines produce → owners receive income → government taxes economic production → government distributes some of the proceeds → people consume and live without necessarily holding traditional jobs.

At that point, Social Security would no longer really be a "retirement insurance system" in the traditional sense. It would become part of a broader system for distributing the enormous productivity generated by automated capital. The transition could be more dangerous than the final state

This is probably the most important distinction.
A fully automated economy could theoretically be extremely
prosperous.

The dangerous period is the transition.
118. For example:

Phase 1:
-----------
AI makes workers more productive
→ wages and profits rise.

Phase 2:
----------
AI replaces some occupations → displaced workers move into other occupations.

Phase 3:
-----------
AI becomes capable of performing increasingly general cognitive and physical work → labor demand begins falling.

Phase 4:
-----------
capital ownership becomes increasingly important→ wages become a smaller share of national income.

Phase 5:
-----------
government redesigns taxation and social insurance → society develops a new mechanism for distributing automated production.

The fiscal system can probably handle phases 1 and 2.

Phases 3–5 would require structural changes.
The paradox
-----------------
There is a fascinating paradox here:

AI could simultaneously make America richer and make the existing social-security financing mechanism weaker.

That's because:
GDP measures how much the economy produces.
while:
payroll taxes depend heavily on how much humans earn from working.

Those two things could increasingly diverge.

CBO's current assessment is appropriately cautious: the overall fiscal impact of AI is uncertain, because productivity and new economic activity could increase revenues while labor displacement could reduce wage and payroll taxes and increase government spending. According to the Congressional Budget Office

And that leads to what I think is the central policy question:

If AI eventually allows 100 people to produce what previously required 200 people, should the U.S. continue financing Social Security, Medicare and the social safety net primarily by taxing the 100 people who still work—or should it begin taxing the machines, capital, profits, consumption, or accumulated wealth that make the higher productivity possible?

That is ultimately less a technology question than a question about who owns the productive capacity of the future and how society chooses to distribute its benefits.

This is, at best a very sticky wicket

Seems like there will be some self-limiting aspect to all of this. If AI actually results in huge job losses... who is going to buy all that cheap stuff AI/robots are producing? And I don't think the population will just sit still and watch this all happen, there will be huge political pressure to prevent seismic disruption. But it could certainly cause a lot of problems.

In the end, I'm not fully convinced this will be any different from previous big changes, such as the industrial revolution, where machines started doing many of the things that humans traditionally did. That created a lot of social problems along the way, but do you want to argue that it would be better to have more ditch-digging jobs and fewer backhoes? Then the automobile made some very profound changes to society, to the point of physically shaping cities into sprawling suburbs. And car crashes kill over 40,000 people every year.

You could make a convincing argument that this has all been bad for mankind, especially considering the pollution and destruction of the environment. But do you want to go back to walking and riding horses? I suppose such an argument can be made. But this is what usually passes for "progress" as we define it today. And it is very hard to stop... no matter what the bad effects are.

Think about this if a machine sees a crash is coming and there is no way to avoid it has to choose between hitting your family and hitting a school bus then you know who is going to die that day.

When I first started in computers / data-processing & programming in the 70's ( long before my involvement with AI ), I became innately aware automation could & would increase the profitability for corporations, while having a high probability of displacing certain jobs fulfilled by humans over time.

At that time, I pondered how many people would need to be displaced by automation systems before some of the social programs supported by income taxes would be impacted.

My peers and I would have long philosophical conversations about what jobs could be overtaken by "automation" and what could be the effect to our society. We easily predicted the self-check out systems. While an individual checkout station might not create a meaningful loss in tax revenue, when combined en mass, a financial tidal wave might be plausible.

In the 70's & early 80's we'd discuss the pie in the sky ideas of true, self aware automation and robotic systems.

Today, robotics are little more than "glorified automation" programs / systems that cooperate and interact together because of some external stimulus. If theres no external stimulus that causes the automation systems to perform / take a specific action, that automation system sits and collects ( what I've termed ) "electronic dust."
As these automation systems start to receive "external stimulus" from other automation systems, and they in turn operate in harmony completing tasks and displacing human jobs, the financial impact to the social services provided by a government is likely to become an unsurvivable financial tsunami.

Is this "progress?" I suppose it depends upon ones point of view.

From my perspective IFF ( thats IF and only IF ) we somehow manage to create truely "self-aware" machines capable of teaching and learning from each other, we'll be on the precipitous of financial ruin.

Think of HAL in 2001 A Space Odyssey... but think of it this way..

Airplanes could be designed and manufactured by robots in such a way they could be controlled by self-aware air traffic vontrol systems. The plane could be self-aware to know its normal maintence schedule as well as when there were anomalous conditions requiring unscheduled maintenence.

The list of possibilities becomes literally endless...

So I'll pose the query again..

Is this progress?

I dunno.. I truly dont have the requisite brain capacity, intelligence, forward vision nor grey brain matter to sort it out!

But I'll agree, I'm not convinced we can stop this "Progress" freight train.