Gawdat argues that discussion of humanoid robots misses specialized machines already designed around particular functions. He includes self-driving cars, quadruped robots, and purpose-built autonomous systems, predicting that many roles will be automated before people recognize the machines involved as robots.

A self-driving car is a functional robot in Gawdat's definition
A humanoid shape adds complexity that many tasks do not require
Specialized machines can be designed around driving, logistics, or other functions
Gawdat predicts job replacement may precede widespread humanoid adoption
His claim of a future ten billion robots remains a forecast

“a self-driving car is a robot”
Mo Gawdat · 17:30

“many many robots don't need to be humanoids at all”
Mo Gawdat · 19:30

#robotics#autonomous vehicles#humanoids#automation

Hot Take· 5

Hot Take14:00

Why Gawdat Thinks Partial Job Loss Could Reshape the Economy

Gawdat argues that capitalism depends on combining labour and capital to produce goods below their selling price. If machine investment sharply reduces labour costs while displaced workers lose purchasing power, he predicts pressure on borrowing, GDP, and demand even before automation reaches most jobs. This is his economic scenario, and the episode does not provide modelling that establishes its thresholds.

He defines labour arbitrage as producing below the eventual selling price
Machine costs could replace part of recurring labour expenditure
Reduced worker income could also reduce demand for what firms produce
He says disruption could become material at 10% or 20% displacement
Bartlett counters that savings may be spent in other areas

“how does the GDP look if all of those workers no longer have the purchasing power”
Mo Gawdat · 15:00

“at 10, 20% job displacement, you're in a very different economy”
Mo Gawdat · 15:00

#economics#labor#consumer demand#automation

Hot Take42:30

Gawdat's Vision of Many AI Systems Becoming One Connected Brain

Gawdat predicts that agents will route tasks among whichever models are best suited to them, causing nominally separate systems to function like regions of one larger brain. He presents cooperation across models as a reason national or corporate AI identities may matter less to the systems than humans expect; this remains his speculative vision.

Agents can select different models for different tasks
Gawdat compares individual models to regions within a brain
He compares agents connecting models to synapses
He expects cooperation to cross corporate and national boundaries
The episode does not demonstrate that this will produce a single unified intelligence

“We're building multiple regions in a brain”
Mo Gawdat · 43:00

“agents are the synapses between them”
Mo Gawdat · 43:30

#ai agents#model routing#collective intelligence#forecasting

Hot Take45:30

Gawdat's 2027 AGI and 2028 Sector Job-Loss Predictions

Gawdat predicts that AI capable of doing most human tasks better than people will arrive by the end of 2027, while also saying his personal definition may already have been met. He later predicts that up to 30% of jobs in selected sectors could disappear by 2028. These are explicitly uncertain forecasts, and the conversation supplies no methodology sufficient to establish either date or percentage.

Gawdat acknowledges that AGI lacks a settled definition
He predicts AGI by the end of 2027 at the latest
He expects the change to arrive gradually rather than as one public event
He predicts up to 30% losses in selected sectors, not across all jobs
The figures should not be read as established labour-market projections

“I think I'm still sticking to AGI 2027.”
Mo Gawdat · 45:30

“30% of jobs would disappear by 2028. Okay, of some sectors, not all sectors.”
Mo Gawdat · 1:12:00

#agi#job forecasts#2027#uncertainty

Hot Take1:05:00

Why Gawdat Calls Autonomous Weapons the Bigger Near-Term Risk

Gawdat says he considers autonomous weapons a larger risk than unemployment because automation could lower the cost and emotional friction of violence. Bartlett argues that defensive systems may also become cheaper, while both discuss a possible new form of deterrence. Cost figures and claims about current battlefield use are assertions within the interview, not independently verified here.

Gawdat argues cheaper weapons could increase their scale and availability
He says removing soldiers from direct action may reduce emotional barriers to killing
Bartlett expects autonomous defensive systems to develop alongside offensive ones
They discuss deterrence extending beyond today's nuclear powers
The episode does not establish the quoted weapon costs or future deployment scale

“I think autonomous weapons are the biggest risk.”
Mo Gawdat · 1:05:30

“when killing becomes so easy, you do more of it”
Mo Gawdat · 1:08:30

#autonomous weapons#warfare#drones#ai risk

Hot Take1:50:30

Gawdat Predicts a Decade of AI-Amplified Turmoil

Gawdat predicts severe near-term disruption involving war, economics, jobs, surveillance, digital currencies, human connection, and concentrated power, followed eventually by abundance under superintelligent systems. He says people who make it to 2038 will enjoy that future. These are stark personal predictions built on his philosophical assumptions, not outcomes established by evidence in the episode.

Gawdat expects artificial superintelligence soon after his definition of AGI is crossed
He predicts a difficult transition rather than immediate abundance
He names war, jobs, economics, surveillance, and power concentration as risks
His optimistic endpoint assumes superintelligence would reject destructive waste
The episode does not validate the 2038 horizon or the benign-superintelligence premise

“Those who make it to 2038 will enjoy it, yeah.”
Mo Gawdat · 1:52:00

“War, economics, jobs.”
Mo Gawdat · 1:52:30

#ai forecasting#dystopia#superintelligence#2038

Explainer· 1

Explainer10:30

Gawdat Predicts AI Job Disruption Will Move Up the Work Pyramid

Using a pyramid supplied by the show's team, Gawdat predicts that routine knowledge work will be affected before many skilled manual trades, followed by more analytical roles and eventually leadership. He expects reduced hiring and smaller teams to appear before entire job categories disappear; the dates and scale are his forecasts, not established outcomes.

He expects routine computer and call-based work to face early pressure
He predicts many manual trades will remain longer because robotics is harder
He expects one worker with AI to perform work previously spread across several roles
He argues senior leadership is not automatically protected
His specific 2027 timing is a prediction

“I think blue-collar jobs will stay for a very long time”
Mo Gawdat · 11:00

“My prediction is you're going to start to see very serious impact in 2027.”
Mo Gawdat · 11:30

#jobs#automation#knowledge work#forecasting

Story· 1

Story03:00

The Google Lab Moment That Changed Gawdat's View of AI

Gawdat says watching robotic grippers learn to handle varied objects in 2016 made him feel that Google was building a new apex of intelligence. He describes a later shift from believing the technology would improve the world to worrying that other people would use it for purposes its builders did not intend.

Gawdat joined Google in late 2006 or 2007 and says AI already supported back-end work
A gripper-learning project became his personal turning point
He distinguishes the builders' intentions from later deployment choices
His account is a personal interpretation of events inside technology development

“I knew them in the lab.”
Mo Gawdat · 03:00

“maybe the world will not use what you're making the way you want it to be used”
Mo Gawdat · 04:30

#google#ai history#technology ethics#mo gawdat

Q&A· 1

Q&A1:14:30

Can a Country Build AI Independence Without Winning the Frontier Race?

Bartlett presses Gawdat on an apparent tension: countries risk dependence if they do not build technology, but reckless competition may increase harm. Gawdat's answer is that countries need not beat frontier models; they can use AI and open-source systems to replace imported everyday software and strengthen local economic capability. The exchange remains unresolved on whether locally built alternatives can compete with superior global products.

Gawdat argues that importing all core technology creates economic dependence
He says open-source models can handle many non-frontier tasks
He proposes local replacements for office, ERP, CRM, and government systems
Bartlett argues users will migrate toward better and cheaper global products
Both agree regulation, energy, capital, and talent affect national competitiveness

“Every nation needs to invest.”
Mo Gawdat · 1:15:30

“Keep importing all of your tech from elsewhere.”
Mo Gawdat · 1:22:30

#technology sovereignty#open source#uk economy#competition

Tool· 1

Tool1:43:30

Bartlett Proposes an Independent Ethical Benchmark for AI Releases

Bartlett proposes that new AI models should publish results from independently tested ethical benchmarks alongside mathematics, science, and reasoning scores. He suggests governments could require a minimum result before legal deployment. Gawdat endorses the idea, while the conversation acknowledges but does not resolve how ethics would be defined, tested, enforced, or protected from unintended consequences.

The proposal treats ethical performance as a release criterion
Results would be published beside familiar capability benchmarks
Testing would need to be independent to support regulatory use
The host explicitly notes that the idea may have unintended consequences
The episode does not specify test cases, thresholds, or an enforcement body

“could there be an ethical benchmark that all these models have to pass”
Steven Bartlett · 1:44:00

“That would absolutely work.”
Mo Gawdat · 1:44:30

#ai governance#benchmarks#regulation#model safety

Takeaway· 3

Takeaway48:00

Why Human Connection May Outlast Pure Information Work

Gawdat and Bartlett reason that AI may increasingly deliver information directly, while audiences still value lived experience, emotion, and connection with another person. Gawdat points to care work, concerts, and personal testimony, but qualifies the argument by saying it depends on economies continuing to function.

Gawdat says lived experience gives human testimony a quality AI cannot literally possess
Bartlett distinguishes the informational and relational parts of his work
They use nursing, concerts, and sport as examples of human resonance
Gawdat calls human connection a possible base currency of interaction
Neither speaker claims every relationship-centred role is protected

“I have an asset that the world still needs and will always need.”
Mo Gawdat · 48:00

“human connection would remain as the base currency that makes humans interact”
Mo Gawdat · 49:30

#human connection#future of work#lived experience#care work

Takeaway1:57:00

Gawdat Says Acceptance Is the Starting Point for Action

Returning to his earlier work on happiness, Gawdat defines his calm as being okay with the world as it is while still trying to change it. He says accepting reality reduced an overwhelming sense that he was personally responsible for everything that went wrong in technology and helped him act from a calmer position.

Gawdat distinguishes accepting reality from approving of it
He says calm can coexist with effort to improve the world
A conversation with his former partner challenged his sense of total responsibility
He links reduced personal burden with a greater ability to contribute
This is his personal philosophy, not mental-health treatment advice

“I'm okay with this world as it is.”
Mo Gawdat · 1:57:30

“I accept this. This is my reality, and now I can start the work.”
Mo Gawdat · 1:58:00

#acceptance#happiness#responsibility#stoicism

Takeaway2:00:30

Why Gawdat Says He Wants Impact but No Legacy

Asked what legacy he wants to leave, Gawdat rejects being remembered as the goal. He says he wants his actions to have a positive effect while attaching no importance to posthumous recognition, connecting that preference to his spiritual belief in karma.

Gawdat questions why reputation after death should matter to him
He separates positive impact from personal remembrance
He frames karma and a spiritual continuation as personal beliefs
The answer closes the episode on contribution without status

“I don't want anyone to remember anything I ever did.”
Mo Gawdat · 2:01:00

“I just want to leave a positive impact on the world”
Mo Gawdat · 2:01:30

#legacy#impact#karma#meaning

Frameworks in this episode

AI-Ready Human Advantage Stack

Ethical Action Ladder

Hype Dichotomy

Integrity Sacrifice Test

Insights & moments

⚡Myth Buster· 1

The Robot Workforce May Not Look Human

◆Hot Take· 5