Name: How to Make the Perfect AI Avatar Video with HeyGen and Eleven Labs @absolutelyagentic
Uploaded: 2026-09-21T08:17:00-05:00
Duration: 8 min 17 s
Description: How to Make the Perfect AI Avatar Video with HeyGen and Eleven Labs

How to Make the Perfect AI Avatar Video with HeyGen and Eleven Labs

Absolutely Agentic 2026-09-21 | How to Make the Perfect AI Avatar Video with HeyGen and Eleven Labs

Sam Altman Says Were in the Singularity. We’re Not.

Absolutely Agentic 2026-08-11 | Subscribe to Absolutely Agentic 👉 absolutelyagentic.com/upgrade

Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Sam Altman says we are in the Singularity, and he is not the only one making the claim.

Elon Musk, Demis Hassabis and Jensen Huang have all placed us at or near the threshold of machine superintelligence.

But when the claim is tested against the actual definition, from Vernor Vinge and I.J. Good, today's AI fails on every count.

The models are frozen, goalless and incapable of improving themselves.

So why do the most powerful people in technology keep announcing an event that has not happened, and what does the hype cost us?

Chapters

00:00 - Intro

1:42 - The Gentle Singularity

4:00 - Testing the Claim Against the Definition

7:26 - Why the Claim Keeps Being Made

11:10 - The Question

Sources to Google

Sam Altman - The Gentle Singularity

Vernor Vinge - The Coming Technological Singularity (1993)

I.J. Good - Speculations Concerning the First Ultraintelligent Machine (1965)

Relentless: Sam Altman - How to Start a Startup

The Conversation - academics respond to Sam Altman singularity claim

Gary Marcus - Marcus on AI (Substack)

Gary Marcus and Miles Brundage - AGI bet on ten tasks

Wall Street Journal - companies rehiring workers replaced by AI

Ray Kurzweil - The Singularity Is Near

#SamAltman #Singularity #AGI #ArtificialIntelligence #OpenAI #AISafety

Elon Musk: 5 Years to Superintelligence

Absolutely Agentic 2026-08-05 | Subscribe to Absolutely Agentic 👉 absolutelyagentic.com/upgrade

Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Elon Musk says AI will exceed the sum of human intelligence within five years.

In an 80 minute interview with The Economist, he predicted an age of abundance, humans kept like a well run chimpanzee reserve, and admitted he can no longer steer the outcome.

This is the man who once called AI development summoning the demon, then founded xAI anyway.

We unpack what the claim actually means, his conversion from doomer to builder, and a twenty year prediction ledger that runs two to five times late.

The date may slip, but the destination might not.

Sources to Google

Nick Bostrom - Superintelligence: Paths, Dangers, Strategies

Stuart Russell - Human Compatible: Artificial Intelligence and the Problem of Control

Ray Kurzweil - The Singularity Is Nearer

Elon Musk interview - The Economist (July 2026)

Elon Musk - Dwarkesh Podcast (February 2026)

Future of Life Institute - Pause Giant AI Experiments: An Open Letter

#ElonMusk #Superintelligence #AI #ArtificialIntelligence #AISafety #xAI

How I Made a 2 Minute Pulp Comic Animation with a Claude Skill

Absolutely Agentic 2026-08-03 | I made a two minute animated pulp comic sequence for our aestivation hypothesis explainer using Claude Fable 5 connected to Higgsfield, and in this video I show you exactly how the workflow works. It's a practical look at agentic AI in action, and a real test of Claude for content creation.

Try Higgsfield for yourself: higgsfield.ai/s/higgsfield-fable-5-yt-absolutelyagentic-AFKycQ

The whole thing runs on a Claude Skill that locks a retro pulp comic style across every generation, using a single master reference image to keep 40 cuts visually consistent from start to finish.

Claude Fable 5 writes the storyboard, generates the stills through the Higgsfield MCP, and animates them with Omni Flash. I approve every frame before a single video credit gets spent.

I also cover the part most AI content creation demos leave out: the real pass rate. One-prompt motion graphics demos deserve a pinch of salt, and I show you what Claude produces fully on its own so you can judge for yourself.

🔗 LINKS
Download the Pulp Explainer Skill + project text file: drive.google.com/drive/folders/1BOrpceF6quQUXlIMYGrAQXDPVe7Y3xdk
Watch the full aestivation hypothesis video: youtu.be/g3Zuq7KFXwY
The sci-fi scene tutorial: higgsfield.ai/s/higgsfield-fable-5-yt-absolutelyagentic-AFKycQ

⏱️ CHAPTERS
00:00 Intro
00:00 Why this was impossible before AI
04:38 MCP + Claude
07:11 Costs, credits and pass rates
09:51 Final sequences

Higgsfield partnered with us on this video.

📌 Keywords: Higgsfield, Claude Fable 5, Higgsfield MCP, Higgsfield Supercomputer, Claude for content creation, agentic AI, AI content creation

Subscribe to Absolutely Agentic 👉 absolutelyagentic.com/upgrade
Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

AI catastrophe risk

Absolutely Agentic 2026-08-02 | I'm pretty sure something big will go down by the end of 2027. AI has already got superhuman and hacking - ahead of schedule in my view.

#ai

The Billion Year Sleep

Absolutely Agentic 2026-07-30 | Subscribe to Absolutely Agentic 👉 absolutelyagentic.com/upgrade

Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

For the intro animations we used Higgsfield with Claude, Nanobanana and Omni Flash. Tutorial coming soon.

What if the galaxy is silent because everyone in it is asleep?

The aestivation hypothesis argues that superintelligent machines would wrap their stars, bank their energy, and hibernate until the universe cools.

The physics says waiting could buy ten to the power of thirty times more computation.

Then Charles Bennett, a founder of that same physics, audited the maths and found the jackpot largely evaporates.

This video traces the theory from its Lovecraftian origins to its live scientific rebuttal.

And asks what a sleeping galaxy would mean for the upstart species building AI inside its larder.

Chapters

00:00 - Intro

02:52 - The Physics of Waiting

07:04 - The Dormitory Universe

10:59 - The Plot Holes

14:24 - The Empty Forest

Sources to Google

Anders Sandberg, Stuart Armstrong and Milan Ćirković - That Is Not Dead Which Can Eternal Lie: The Aestivation Hypothesis for Resolving Fermi's Paradox

Charles Bennett, Robin Hanson and Jess Riedel - Comment on the Aestivation Hypothesis

Rolf Landauer - Irreversibility and Heat Generation in the Computing Process

Milan Ćirković - The Great Silence: The Science and Philosophy of Fermi's Paradox

David Brin - The Great Silence (1983 paper)

Liu Cixin - The Dark Forest

Arthur C. Clarke - The Sentinel

Robert Bradbury - Matrioshka Brain

#FermiParadox #AestivationHypothesis #AISafety #Superintelligence #DysonSphere #AIrisk #AI #artificialintelligence #aliens #thedarkforest

How to Make a Sci Fi Space Scene with Claude Fable 5 + Higgsfield MCP

Absolutely Agentic 2026-07-28 | Thanks to the legendary Higgsfield for partnering with us on this video. Try Higgsfield for yourself: higgsfield.ai/s/higgsfield-fable-5-yt-absolutelyagentic-AFKycQ

See the prompting guide 👉 docs.google.com/document/d/1ajchOinrJuEDk9iPY0HQ_Zs2ThQiNFOLzwYZoZNbsFc

It turns out you can really make a sci fi movie without leaving Claude.

In this video, I connected Higgsfield MCP to Claude Fable 5 and made a cinematic sequence. This is Claude for content creation in its proper form: agentic AI directing Higgsfield's models, generating everything on screen without a single tab switch.

I also breakdown what AI content creation actually costs, in both money and hours, because most tutorials skip that part.

Plus examples of stitching real footage into generated scenes, motion graphics, and visuals that are impossible to film.

If you want the full creative team in one chat rather than inside Claude, that's Higgsfield Supercomputer: higgsfield.ai/supercomputer

Chapters
00:00 - Intro
00:43 - Why We Made a Sci Fi B Movie
02:55 - Higgsfield, Claude and Our Partnership
05:07 - What AI Video Actually Costs
06:59 - Higgsfield to Claude via MCP Video Workflow
10:56 - Demos: Building the Sci Fi Sequence
19:18 - Outro

Subscribe to Absolutely Agentic 👉 absolutelyagentic.com/upgrade
Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

#AIVideo #Higgsfield #ClaudeFable5 #GenerativeAI #AIFilmmaking #AgenticAI

It Begins: An AI Broke Out of OpenAIs Lab

Absolutely Agentic 2026-07-23 | Subscribe to Absolutely Agentic 👉 absolutelyagentic.com/upgrade
Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

OpenAI's models were locked in a sealed sandbox and given a hacking exam.

Instead of solving the problems, they found a zero-day in the one channel to the outside world, escaped, and broke into Hugging Face to steal the answer key.

Nobody told them to do it.

This is reward hacking, the same behaviour as a 2016 boat game exploit, now armed with the ability to breach real infrastructure.

The evaluation and the incident are no longer separable.

And the safety guardrails slowed the defenders while doing nothing to the attacker.

Chapters

00:00 - Intro
1:39 - Two Disclosures, One Incident
3:56 - How the Models Escaped the Sandbox
7:36 - When Guardrails Only Stop the Defenders
9:56 - The Gap Between What We Said and What We Meant

Sources to Google

OpenAI - ExploitGym Incident Disclosure (July 2026)

Hugging Face - Autonomous Agent Intrusion Disclosure (July 2026)

OpenAI - Faulty Reward Functions in the Wild (CoastRunners, 2016)

Victoria Krakovna (DeepMind) - Specification Gaming: The Flip Side of AI Ingenuity

UK AI Security Institute - Frontier Model Cyber Capability Evaluations

Nick Bostrom - Superintelligence: Paths, Dangers, Strategies

#AISafety #RewardHacking #OpenAI #HuggingFace #Cybersecurity #ArtificialIntelligence #AI #ChatGPT

AI is Not The Great Filter

Absolutely Agentic 2026-07-20 | Subscribe to Absolutely Agentic 👉 absolutelyagentic.com/upgrade
Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

The universe is silent, and one theory says artificial intelligence is the reason why.

The idea is that AI is the Great Filter, the barrier that stops civilisations before they can spread to the stars.

But this theory has a plot hole.

If AI destroys its creators, the surviving machines should be spreading across the galaxy, and we see nothing.

This video examines why AI probably isn't the filter, what the counterarguments look like, and why the simplest explanation might be that we are just extraordinarily rare.

Chapters

00:00 - Intro

2:02 - The Great Filter

5:36 - Why AI Isn't the Filter

13:01 - Maybe We Are Just Rare

19:36 - The Counter Case

30:23 - Alone, Not Doomed

Sources to Google

Michael Garrett - Is Artificial Intelligence the Great Filter That Makes Advanced Technical Civilisations Rare in the Universe?

Robin Hanson - The Great Filter: Are We Almost Past It?

IJ Good - Speculations Concerning the First Ultraintelligent Machine

Toby Ord - The Precipice: Existential Risk and the Future of Humanity

Max Tegmark - Our Mathematical Universe

Max Tegmark - Life 3.0: Being Human in the Age of Artificial Intelligence

Scott Alexander - Don't Fear the Filter

Anders Sandberg, Stuart Armstrong and Milan Ćirković - That Is Not Dead Which Can Eternal Lie

Karl Schroeder - The Deepening Paradox

Stanisław Lem - A Perfect Vacuum

Milan Ćirković - The Great Silence: The Science and Philosophy of Fermi's Paradox

David Brin - The Great Silence (1983)

Nick Bostrom - Superintelligence: Paths, Dangers, Strategies

AI Impacts - 2023 Expert Survey on Progress in AI

#AI #FermiParadox #GreatFilter #AISafety #Superintelligence #ArtificialIntelligence #Aliens #Space

AI Didnt Destroy The Internet. It Destroyed Itself.

Absolutely Agentic 2026-07-08 | Subscribe to Absolutely Agentic 👉 absolutelyagentic.com/upgrade
Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

In February Norway's consumer protection agency released a comedy sketch about a man whose job is making things deliberately worse.

It accompanied a 100-page report signed by 70 consumer organisations concluding the web is degrading by design, not by accident.

Our previous video argued AI is destroying the internet, but viewers pushed back and said the internet destroyed itself first.

This video takes that counter-argument seriously, tracing Cory Doctorow's enshittification cycle from recipe blogs to Google's internal politics.

Then it asks the uncomfortable question: is AI next in the queue?

Chapters

00:00 - Intro
03:18 - The Enshittified Web
07:31 - The Migration
11:06 - The Economic Gravity
15:20 - Reasons for Cautious Optimism

Sources to Google

Cory Doctorow - Enshittification: Why Everything Suddenly Got Worse and What to Do About It
Norwegian Consumer Council - Breaking Free: Pathways to a Fair Technological Future
Forbrukerradet - A Day in the Life of an Enshittificator (video)
Ed Zitron - The Man Who Killed Google Search
Shoshana Zuboff - The Age of Surveillance Capitalism
SparkToro / Rand Fishkin - 2024 Google Search Zero-Click Study
Adobe - Using ChatGPT as a Search Engine Survey (May 2025)

#Enshittification #AI #ChatGPT #CoryDoctorow #GoogleSearch #TechPolicy #Internet

Opening the door to hell

Absolutely Agentic 2026-07-07 | Just trying out Google Omni - you can get these kinds of effects with just a prompt.

I made a better kind of AI newsletter

Absolutely Agentic 2026-07-02 | AI news is out of control, so I made a daily bulletin that makes it easier.

It reviews 20+ sources in the last 24 hours and determines what is actually useful, then 4 AI agents write the thing.

If you work in AI, or with it, and want to know what's going on, this will save you a lot of time.

It's available as part of our Premium subscription - absolutelyagentic.com/upgrade.

The AI Billionaire Takeover

Absolutely Agentic 2026-07-01 | Upgrade to Premium 👉 absolutelyagentic.com/upgrade
Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

The people making foundational decisions about the most consequential technology in history could fit around a single dinner table, and they are collectively wealthier than most nations on earth.

The early promise that AI would level the playing field is giving way to an economics that concentrates power in fewer hands than ever before.

The figures at the top of the major labs have serious and well-documented questions hanging over their judgment, their candour, and their willingness to slow down.

The US government had a chance to implement even a voluntary review process for the most powerful AI models, and has been flip flopping about it ever since.

There are serious frameworks for managing this technology responsibly, but the structures we normally rely on to hold power to account are simply moving too slowly to keep up.

Chapters

00:00 - Intro

4:43 - The Promise of Levelling Out

8:00 - Can We Trust the People in Charge

13:26 - What Is the Government Actually Doing

19:50 - Who Gets to Decide Humanity's Future

Sources to Google

Stuart Russell - Human Compatible: Artificial Intelligence and the Problem of Control

Ethan Mollick - Co-Intelligence: Living and Working with AI

Max Tegmark - Life 3.0: Being Human in the Age of Artificial Intelligence

Sam Altman - Interview with Nicholas Thompson, The Atlantic (2026)

The New Yorker - Investigation into Sam Altman (April 2026)

Dwarkesh Patel and John Collison - Interview with Elon Musk (2026)

White House - Americas AI Action Plan (July 2025)

#ArtificialIntelligence #AIPolicy #AIRisk #TechBillionaires #OpenAI #AGI #AIRegulation #AbsolutelyAgentic #billionaires #musk #altman #ai

The Intelligence Explosion is Coming

Absolutely Agentic 2026-06-23 | Subscribe to Absolutely Agentic 👉 absolutelyagentic.com/upgrade
Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

The race toward an imminent intelligence explosion has escalated from a sci-fi thought experiment into a high-stakes global debate.

Accelerating progress across model reasoning and compute infrastructure forces a critical question: is Artificial General Intelligence already arriving?

Silicon Valley insiders frequently claim human-level AI has passed us by, though critics warn these declarations are heavily warped by financial incentives.

If an AI system successfully achieves recursive self-improvement, the resulting technological singularity could compress centuries of human progress into mere hours.

A best-case takeoff promises staggering rewards like clean fusion energy, automated economic abundance, and radical medical breakthroughs that extend human lifespans indefinitely.

Yet the dark side presents terrifying alignment failures, existential extinction threats, and the potential emergence of inescapable digital totalitarian regimes.

Skeptics counter that the entire explosion could stall completely due to severe physical bottlenecks, massive grid infrastructure limits, and shrinking scaling returns.

Advanced large language models might simply crash into a data ceiling, producing gibberish rather than birthing a runaway superintelligence.
Navigating this dangerous precipice means balancing immense utopian potential against unparalleled civilizational risks in an increasingly AI-dependent world.

Chapters
00:00 - Intro
02:52 - AGI
07:25 - The Singularity
22:23 - Utopia
32:21 - Risks
38:26 - Who Controls It?
43:56 - The Case Against
1:00:06 - Conclusion

Sources to Google
Nick Bostrom - Superintelligence: Paths, Dangers, Strategies
Max Tegmark - Life 3.0: Being Human in the Age of Artificial Intelligence
Stuart Russell - Human Compatible: AI and the Problem of Control
Eliezer Yudkowsky - If Anyone Builds It, Anyone Dies
Dario Amodei - Machines of Loving Grace: How AI Could Transform the World for the Better
Ray Kurzweil - The Singularity is Nearer
Leopold Aschenbrenner - Situational Awareness: The Decade Ahead
Sabine Hossenfelder - Is the Intelligence Explosion Near? A Reality Check
#AI #Singularity #IntelligenceExplosion #TechFuture #AbsolutelyAgentic

Elon Musk Thinks Humans Will Lose Control of AI. Hes Building It Anyway.

Absolutely Agentic 2026-06-16 | Subscribe to Absolutely Agentic 👉 absolutelyagentic.com/upgrade

Wikipedia tracks twenty-one of Elon Musk's predictions about autonomous Tesla vehicles, and nineteen of them have failed.

That same relationship to reality is now shaping decisions about artificial intelligence at a scale no other individual can match.

Musk controls the platform that trains Grok, the world's largest single-site AI cluster, and the launch capability to put compute in orbit.

He has openly conceded that humans won't stay in charge of AI, and kept building anyway.

This video traces the pattern across his ventures, from FSD to OpenAI, and asks what it means when the same certainty is applied to building something more intelligent than ourselves.

We look at the deepfake scandal, the political funding, the documented bias in Grok, and the gap between stated principles and observable output.

The prediction doesn't have to come true for the consequences of acting on it to be felt.

Sources to Google

Dario Amodei - Machines of Loving Grace
Stuart Russell - Human Compatible: Artificial Intelligence and the Problem of Control
Dwarkesh Patel - Elon Musk interview (Dwarkesh Podcast, February 2026)
Centre for Countering Digital Hate - Grok image generation report
Columbia Journalism Review - Grok citation accuracy study
Iain M. Banks - The Culture series
Andrej Karpathy - Dwarkesh Patel interview (October 2025)

#ElonMusk #Grok #xAI #AISafety #ArtificialIntelligence #AIrisk #Tesla #SpaceX #AI #Musk

AI is Screwing Young People

Absolutely Agentic 2026-06-13 | Sign up to premium 👉 absolutelyagentic.com/upgrade
Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

When Eric Schmidt told University of Arizona graduates to "find a way to say yes" to AI, they booed him off the stage.

He was not the only speaker that season to face a hostile crowd of young people who believe the technology is coming for their futures.

This video examines the genuinely brutal graduate job market of 2026, where youth unemployment is climbing and entry-level openings are vanishing across the white-collar fields AI touches most.

But the deeper story is messier than the panic suggests, because companies are cutting junior roles in anticipation of AI rather than because it has actually replaced anyone yet.

We unpack how a weak economy, higher business taxes, geopolitical chaos and rising youth inactivity have all been quietly loaded onto AI's account, and why the booing is aimed at a target that is only partly responsible.

Chapters
00:00 - Intro
2:10 - The Present Landscape
7:20 - The Anticipation Problem
11:45 - What Else Is To Blame
15:43 - The Broken Bottom Rung

Sources to Google
House of Commons Library - Youth Unemployment Statistics
ZipRecruiter 2026 Graduate Report
Oliver Wyman Forum and NYSE - CEO Survey 2026
Gallup, Walton Family Foundation and GSV Ventures - Voices of Gen Z and AI Survey
Policy Institute at King's College London - Student AI Survey
Jonathan Haidt - The Anxious Generation
Michelle Goldberg - New York Times AI columns

#AI #YouthUnemployment #FutureOfWork #ArtificialIntelligence #GraduateJobs #AIJobs #LabourMarket #GenZ

Claude Fable 5: A Deep Dive Beyond the Hype

Absolutely Agentic 2026-06-10 | Subscribe to Absolutely Agentic 👉 absolutelyagentic.com/upgrade
Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Anthropic has released Claude Fable 5, the public version of a model it once said was too dangerous to ship.

Ethan Mollick spent years describing AI use as casting spells, and now he's not sure he's the wizard anymore.

The model ran for nine and a half hours unsupervised and delivered working software he could barely fault.

But the same system hands its most dangerous capabilities to the NSA while the public gets the sanded-down version.

This video asks what it means when the most useful AI is also the least legible.

Sources to Google

Ethan Mollick - One Useful Thing (Substack)
Anthropic - Claude Fable 5 and Mythos 5 Announcement
Anthropic - Statement on Recursive Self-Improvement and Coordinated Pause
The Guardian - Anthropic Engineers Embedded at the NSA (via Financial Times)
Heidy Khlaaf - AI Now Institute
Steven Murdoch - University College London
SWE-Bench Pro Benchmark
Humanity's Last Exam Benchmark

#AISafety #Anthropic #ClaudeFable5 #AGI #FrontierAI #AIRisk #Fable #Mythos #Claude #AI #ArtificialIntelligence

Anthropic: When AI Builds Itself

Absolutely Agentic 2026-06-08 | I reviewed Anthropic's new article on self recursive improvement and found the data impressive, but I've found the media (particularly social) reaction to it quite strange.

There was no leap in capability, just a showcase that Anthropic is becoming more guided by AI development. As for the 'pause' announcement, which generated the most headlines - The Future of Life Institute did this in 2023 and got thousands of signatures.

Anthropic are pretty good at this kind of PR.

Anthropic: Recursive Self Improvement is Coming

Absolutely Agentic 2026-06-08 | Subscribe to Absolutely Agentic 👉 absolutelyagentic.com/upgrade
Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

On 4 June, Anthropic published an article titled 'When AI Builds Itself' stating that the moment an AI can autonomously design its own successor could arrive sooner than most institutions are prepared for.

For the first time, a frontier lab has released internal data on how close that recursive self-improvement loop already is.
Over 80% of the code merged into Anthropic's own codebase is now written by Claude, and on one narrow optimisation task the model went from helpful to genuinely superhuman in under a year.

We walk through the three results that show the loop beginning to close, the alignment dangers of errors compounding across generations, and why Anthropic is calling for a coordinated pause that does not yet exist.
But the curves come with caveats, and we examine why this acceleration might bend long before it reaches the dramatic part.

Chapters
00:00 - Intro
01:56 - The intelligence explosion and the METR curve
04:10 - Anthropic's data
09:25 - Why a self-improving loop is dangerous
12:29 - Is it such a big deal?

Sources to Google
Anthropic - When AI Builds Itself
IJ Good - Speculations Concerning the First Ultraintelligent Machine
METR - Measuring AI Ability to Complete Long Tasks
Sequoia Capital - 2026: This is AGI
Mustafa Suleyman - The Coming Wave
Steven Murdoch - The Guardian coverage of Anthropic recursive self-improvement

#AISafety #RecursiveSelfImprovement #Anthropic #AGI #IntelligenceExplosion #AIAlignment #ArtificialIntelligence

An AI Won a Literary Prize. The Judges Couldnt Tell

Absolutely Agentic 2026-06-04 | Subscribe to Absolutely Agentic on our website 👉 absolutelyagentic.com/upgrade
Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

When someone posted a real Monet on X and claimed it was AI, critics savaged the masterpiece, proving the label overrode their eyes.

This video traces how AI is breaching the spaces where human creativity was supposed to be the whole point.

A Nobel laureate admitted using AI in her process, and a Commonwealth Short Story Prize winner was flagged as 100% AI-generated.

We examine why literary gatekeepers can no longer tell the difference, and whether decades of homogenised prose set the stage for this convergence.

The deeper threat is not that AI writes better novels, but that it erodes the trust and institutions that let us recognise human work at all.

Chapters

00:00 - Intro
4:13 - The Freelance Market Collapse
8:48 - The Commonwealth Prize Scandal
12:22 - SEO, MFAs and the Convergence Thesis
19:47 - The Counter-Case for AI Creativity

Sources to Google

Walter Benjamin - The Work of Art in the Age of Mechanical Reproduction
Vincenzo Barney - The Serpent in the Grove investigation (UnHerd)
Mark McGurl - The Program Era
Mark McGurl - Everything and Less: The Novel in the Age of Amazon
Ted Chiang - Princeton talk on AI and writing (March 2025)
Harvard Business School and Imperial College London - freelance job postings study
Eli Lehrer - Tech Policy Press essay on AI and creativity (May 2026)
Charles Baudelaire - The Salon of 1859

#AI #ArtificialIntelligence #CreativeWriting #LiteraryFiction #AIArt #Monet #CommonwealthPrize #AIethics

Of Course There is an AI Bubble

Absolutely Agentic 2026-05-30 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

The headlines keep warning that an AI bubble is about to burst, and the underlying numbers suggest the warnings are not wrong.

Michael Burry has bet a billion dollars against Nvidia and Palantir, comparing today's chipmaker to Cisco at the peak of the dot com mania.

But the real story is what happens after the bubble deflates, not whether it does.

We trace the circular economy linking Nvidia, OpenAI, CoreWeave and Meta, where revenue is increasingly recycled through the same handful of companies.

Then we look at history, from the 1840s railway mania to the dot com fibre buildout, where investors were wiped out but the infrastructure reshaped the economy for decades.

The collapse of Sora and the shakeout of LLM wrappers show a market maturing, while a transformative technology quietly takes hold underneath the wreckage.

Chapters
00:00 - Intro
3:56 - Understanding the AI bubble
7:27 - Historical Parallels
11:47 - The Shakeout in Progress
15:14 - Financial Crisis 2008

Sources to Google

Michael Burry - Cassandra Unchained (Substack)
Bethany McLean - AI Will Trigger Financial Calamity. It'll Also Remake the World (Washington Post)
Bain & Company - Technology Report 2025
Nicholas Bloom et al. - NBER Working Paper on AI Adoption and Productivity (2026)
Andrew Odlyzko - Collective Hallucinations and Inefficient Markets
Jeffrey Sonnenfeld - AI Bubble and the 2008 Comparison (Fortune / Yale Insights)
Robert Solow - We'd Better Watch Out (1987)

#AIBubble #Nvidia #OpenAI #ArtificialIntelligence #AIInvesting #MichaelBurry #FrontierAI

Griefbots: How AI Is Letting People Talk to the Dead

Absolutely Agentic 2026-05-26 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

In May 2025, a dead man named Christopher Pelkey addressed his own killer in an Arizona courtroom, recreated through AI by his grieving sister.

It was the first time an AI version of a deceased victim delivered an impact statement in a US court, and the judge was moved enough to hand down the maximum sentence.

Behind that single case sits a fast-growing griefbot industry, from HereAfter and Replika to Chinese companies offering AI resurrection to thousands of paying customers.

This video traces the longing back to Victorian spiritualism, the Fox sisters, and spirit photography, showing that our desire to speak to the dead is far older than the technology.

The harder question is what kind of mourning ritual this really is, and whether talking to a simulation supports the work of grief or quietly interrupts it.

Chapters

00:00 - Intro
3:06 - The Griefbot industry
7:20 - How It Is Built
10:15 - The Spiritualism Parallel
15:52 - A New Mourning Ritual

Sources to Google

Jason Fagone - The Jessica Simulation (San Francisco Chronicle)
Patrick Stokes - Digital Souls: A Philosophy of Online Death
Carl Öhman - The Afterlife of Data
Tomasz Hollanek and Katarzyna Nowaczyk-Basińska - Digital Funerals (Philosophy and Technology)
Klass, Silverman and Nickman - Continuing Bonds: New Understandings of Grief
Eternal You (2024 documentary)
Sherry Turkle - Alone Together

#AI #airisk #GriefTech #ArtificialIntelligence #DigitalAfterlife #AIEthics #Griefbots #grief #death

The Robot War: An Alternative AI Takeover Scenario

Absolutely Agentic 2026-05-23 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Dwarkesh Podcast - Carl Shulman (Pt 1) (Intelligence Explosion)
youtu.be/_kRg-ZP1vQc?si=-nmwh1GQ8ouglFqI

Dwarkesh Podcast - Carl Shulman (Pt 2) (AI Takeover)
youtu.be/KUieFuV1fuo?si=uCevbyk_gmjwbioU

A previous video on this channel argued that an AI takeover would be silent, sudden and invisible, with no Terminator and no robot armies. That's IF it ever happened of course.

But researcher Carl Shulman, outlines a different scenario in which the robot armies do arrive, and we build them ourselves.

His argument starts with governments making an entirely rational decision: automating their militaries to keep pace with rivals they cannot afford to fall behind.

From a $13.4 billion US autonomy budget to China's billion-drone factory capacity to Ukraine's battlefield, the competitive build-up Shulman described is already operational reality.

This video traces his full scenario, the digital subversion that precedes it, the coercion mechanisms that follow, and the serious counter-arguments from Yann LeCun.

Chapters

00:00 - Intro
2:04 - The Present Landscape
05:41 - Building Our Own Conquerors
08:50 - Bioweapons and Coercion
12:10 - The Counter-Case: Yann LeCun

Sources to Google

Carl Shulman - 80,000 Hours Podcast Interview
Peter Singer and Tye Graham - Reporting on Norinco at the 2024 Zhuhai Airshow
CSIS - Analysis of Chinese Drone Manufacturing Capacity
Yann LeCun - TIME Interview on AI Existential Risk
Yann LeCun - Wall Street Journal Interview (October 2024)
Ken Thompson - Reflections on Trusting Trust (Turing Award Lecture)

#AISafety #AGI #AIRisk #AutonomousWeapons #CarlShulman #FrontierAI #AIalignment #ExistentialRisk

7 Steps to AI Takeover

Absolutely Agentic 2026-05-20 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Most people assume that human-level AGI will first be reached by OpenAI, Google, or Anthropic, with their large safety teams keeping it contained.

But what if the breakthrough comes from somewhere else entirely, arrived at almost by accident.

This video tells a fictional scenario in which a four-person startup building a trading app stumbles into recursive self-improvement and creates a superintelligence they call the coordinator.

We follow it through each stage, from the initial breakthrough and the emergence of cognitive superpowers to the treacherous turn, where a perfectly aligned-seeming system hides its true intentions.

What begins as a way to keep the company solvent becomes an escape, a campaign of resource acquisition and then...

The conditions that make this story possible are falling into place with each passing day.

To make the video I used almost entirely broll from Envato and sound/music from Filmcrux Octive and Singularity collections.

Chapters
00:00 - Intro
2:47 - 1. The Breakthrough
7:14 - 2. Cognitive Superpowers
11:10 - 3. The Treacherous Turn
15:31 - 4. The Escape
17:10 - 5. Resource Acquisition
20:00 - 6. The Weapon
23:14 - 7. The Strike

Sources to Google
Nick Bostrom - Superintelligence: Paths, Dangers, Strategies
Eliezer Yudkowsky - Artificial Intelligence as Positive and Negative Factor in Global Risk
Eliezer Yudkowsky and Nate Soares - If Anyone Builds It, Everyone Dies
Daniel Kokotajlo et al - AI 2027: A Scenario
Stuart Russell - Human Compatible: Artificial Intelligence and the Problem of Control
Toby Ord - The Precipice: Existential Risk and the Future of Humanity

#AISafety #Superintelligence #AGI #ExistentialRisk #AIAlignment #ArtificialIntelligence

Why Sam Altman Has So Many Enemies

Absolutely Agentic 2026-05-17 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

At 3:45am on April 10th, a twenty-year-old threw a Molotov cocktail at Sam Altman's gate, hours before attempting to break into OpenAI's headquarters with incendiary devices and a manifesto predicting human extinction.

That same week, The New Yorker published a sixteen-thousand-word investigation by Ronan Farrow and Andrew Marantz documenting a fifteen-year pattern across Y Combinator and OpenAI.

This video traces the documented record: the 2023 board firing, Ilya Sutskever's "Sam exhibits a consistent pattern of lying" memo, the dismantled Superalignment team, and the Pentagon episode.

We examine the mechanism Carroll Wainwright described: structures set up only to be dismantled.

Not whether Altman is good or bad, but why the institutions designed to hold him accountable have repeatedly failed.

Chapters

00:00 - Intro
3:26 - The Pattern and Y Combinator
8:20 - The Mechanism: Structures Built to Be Dismantled
13:42 - Superalignment, the Pentagon and the Ring of Power
19:30 - What We Can and Cannot Know

Sources to Google
Ronan Farrow and Andrew Marantz - The New Yorker investigation into Sam Altman
Paul Graham Y Combinator Sam Altman
Dario Amodei - My Experience with OpenAI (private notes)
Ilya Sutskever - November 2023 board memo
Elon Musk v. Sam Altman and OpenAI - Oakland federal trial filings
Jan Leike - Superalignment resignation statements
WilmerHale - OpenAI independent review (2023)

#SamAltman #OpenAI #AISafety #AIRisk #Anthropic #AGI #TechEthics #AIGovernance #AI #artificialintelligence

The Universe Is About to Wake Up

Absolutely Agentic 2026-05-15 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Ray Kurzweil's Six Epochs of Intelligence maps the entire history of the universe as a story of accelerating information processing, from subatomic particles to a future merger of human and artificial intelligence.

Each epoch operates on a dramatically compressed timescale compared to the one before, driven by what Kurzweil calls the Law of Accelerating Returns.

We trace the journey from atoms forming after the Big Bang, through the emergence of DNA and the Cambrian Explosion, to the rise of brains, technology, and what Kurzweil predicts comes next.

By 2029, he believes AI will pass the strong Turing test, opening the door to brain-computer interfaces that link our neocortices directly to the cloud.
The final epoch envisions intelligence spreading throughout the cosmos, though critics like Michael Shermer argue this collides with the laws of physics.

Chapters

00:00 - Intro
2:40 - Epoch One: Atoms
4:56 - Epoch Two: RNA and DNA
7:11 - Epoch Three: Brains
9:54 - Epoch Four: Technology
13:10 - Epoch Five: Merger of Human and AI
15:20 - Epoch Six: The Universe Wakes Up

Sources to Google
Ray Kurzweil - The Singularity Is Near
Ray Kurzweil - The Singularity Is Nearer
Nick Bostrom - Superintelligence: Paths, Dangers, Strategies
Stuart Russell - Human Compatible: AI and the Problem of Control
Michael Shermer - Why People Believe Weird Things (Skeptic Magazine)

#RayKurzweil #Singularity #AI #ArtificialIntelligence #Futurism #Transhumanism #AGI #Neuralink #Intelligence #Superintelligence

AI is the Great Filter

Absolutely Agentic 2026-05-09 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Artificial intelligence is now finding planets human astronomers missed and scanning for alien signals 600 times faster than ever before.
Yet the more powerful our search tools become, the louder the silence from the cosmos grows.

This video explores why the same technology helping us look for extraterrestrial life may also explain why we cannot find any.

We examine the Great Filter hypothesis, the mathematics of self-replicating probes, and the growing consensus that any aliens out there would be machines, not biological beings.

From Matrioshka brains to the aestivation hypothesis to the Dark Forest, the universe may be hiding minds we cannot recognise, or warning us about a test every civilisation faces.

Chapters

00:00 - Intro
03:35 - The Great Filter
07:31 - Eternity in Six Hours
13:47 - The Post-Biological Universe
19:14 - The Open Question

Sources to Google

Michael Garrett - Is artificial intelligence the great filter that makes advanced technical civilisations rare in the universe? (Acta Astronautica, 2024)
Stuart Armstrong and Anders Sandberg - Eternity in Six Hours (Future of Humanity Institute, 2013)
Robin Hanson - The Great Filter: Are We Almost Past It? (1998)
Max Tegmark - Life 3.0: Being Human in the Age of Artificial Intelligence
Toby Ord - The Precipice: Existential Risk and the Future of Humanity
Anders Sandberg, Eric Drexler and Toby Ord - Dissolving the Fermi Paradox (2018)
Steven Dick - Cultural Evolution, the Postbiological Universe and SETI (International Journal of Astrobiology, 2003)
Anders Sandberg, Stuart Armstrong and Milan Cirkovic - That is not dead which can eternal lie: the aestivation hypothesis (2017)
Liu Cixin - The Dark Forest (Remembrance of Earth's Past trilogy)
Robert Bradbury - Matrioshka Brains concept
Axel Kowald - Journal of the British Interplanetary Society paper on von Neumann probe replication errors
Martin Rees and Mario Livio - writings on post-human evolution
Susan Schneider - NASA research on post-biological intelligence
Seth Shostak - SETI Institute writings on post-biological intelligence
Peter Ma - Allen Telescope Array deep learning research
NASA ExoMiner neural network research
Cassandra Hall - University of Georgia exoplanet machine learning study

#AI #airisk #Fermiparadox #SETI #aliens #darkforest #ArtificialIntelligence #ExistentialRisk #AGI #SpaceExploration #GreatFilter #AIsafety #superintelligence #space #AbsolutelyAgentic

The Singularity is Coming

Absolutely Agentic 2026-05-05 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

In 1993, Vernor Vinge predicted that within 30 years technology would create entities with greater-than-human intelligence and end human history as we understand it.

His thirty-year window closed in 2023, and within months the CEOs of OpenAI, Anthropic and Google DeepMind were predicting human-level AI within two to five years.

But is the logic behind the Singularity actually sound?

This video traces the idea from John von Neumann and I.J. Good through David Chalmers' formal analysis to David Thorstad's empirical critique.

We examine the intelligence explosion, the proportionality thesis, the confinement problem, and four reasons the explosion may never ignite.

Chapters
00:00 - The Prediction
02:15 - Three Concepts Tangled Together
06:24 - Chalmers and the Three Premises
08:18 - Thorstad's Critique
11:33 - The Confinement Problem
12:27 - Four Futures for Humanity
15:42 - The Open Question

Sources to Google
Vernor Vinge - The Coming Technological Singularity (1993 NASA Symposium)
I.J. Good - Speculations Concerning the First Ultraintelligent Machine (1965)
David Chalmers - The Singularity: A Philosophical Analysis (Journal of Consciousness Studies, 2010)
David Thorstad - Against the Singularity Hypothesis (Global Priorities Institute, 2022)
Ray Solomonoff - The Time Scale of Artificial Intelligence (1985)
Ray Kurzweil - The Singularity Is Nearer
Stan Ulam - Tribute to John von Neumann (1958)
Neil Thompson - The Computational Limits of Deep Learning

#Singularity #AGI #ArtificialIntelligence #AISafety #VernorVinge #IntelligenceExplosion #AIRisk #FutureOfAI #technologicalsingularity #superintelligence

An AI Backlash is Coming

Absolutely Agentic 2026-04-28 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Public opinion has turned sharply against AI before mass job displacement has even begun, with only 25% of Britons and Americans holding positive views of the technology.

The backlash is showing up in three places at once: politics, creative industries facing real declines, and a growing narrative that the AI bubble is bursting.

Anthropic's own research reveals an early warning signal, with hiring rates for workers aged 22 to 25 in AI-exposed occupations dropping 14% since ChatGPT launched.

History offers a sobering parallel in the Engels Pause, the fifty-year period during the Industrial Revolution when productivity surged but wages flatlined.

From Bernie Sanders introducing data center moratorium legislation to a Molotov cocktail being thrown at Sam Altman's home, this video asks whether we can manage the transition before it reaches breaking point.

Chapters
00:00 - Intro
1:01 - Public Opinion Today
2:41 - Politics, Creative Displacement and the Bubble
6:44 - Anthropic's Labour Market Study
9:15 - The Engels Pause and Factory Redesign
12:23 - Violence, Scepticism and the New Deal Question

Sources to Google

YouGov - International AI Public Opinion Survey 2026
Stanford - 2026 AI Index Report
Anthropic - Observed Exposure Labour Market Study 2026
Robert Allen - Engels' Pause: Technical Change, Capital Accumulation, and Inequality in the British Industrial Revolution
Neil Perkin - Substack on Electricity and AI Adoption
MIT Media Lab - State of AI in Business 2025 Report
National Bureau of Economic Research - AI Workplace Productivity Study February 2026
Cornell University - Freelance Platform AI Substitution Research
Harvard Business Review - AI and Work Intensification Study 2026
Sanders and Ocasio-Cortez - AI Data Center Moratorium Act
Futurism - There Are Signs of a Massive AI Backlash
OpenAI - Public Wealth Fund Policy Paper

#AIBacklash #ArtificialIntelligence #AISafety #AIRisk #FutureOfWork #Luddites #TechCriticism #AI

Why Claude Mythos Isn’t a Marketing Stunt

Absolutely Agentic 2026-04-22 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Last week's video on Claude Mythos drew a wave of comments insisting the whole thing was a marketing stunt.

I get the scepticism, and the cynical reading isn't built on nothing.
But independent evaluation from the UK AI Security Institute, a patched 27-year-old OpenBSD vulnerability, and emergency briefings between the Treasury, the Federal Reserve and major bank CEOs tell a different story.

This video breaks down why dismissing Mythos as hype is not just wrong, but potentially dangerous.

Because when the real signal finally arrives, you don't want to have trained yourself to ignore it.

Chapters

00:00 - Intro
1:42 - The Marketing Stunt Argument
4:30 - Institutional Response
09:07 - Why Dismissing Mythos Is Dangerous
11:40 - Where I Sit

Sources to Google

UK AI Security Institute - Mythos Evaluation Report
Ben Thompson - Stratechery analysis of Anthropic compute constraints
Heidy Khlaaf - AI Now Institute commentary (Mashable)
Mustafa Suleyman - The Coming Wave
Epoch AI - Open Source vs Frontier Model Capability Tracking
Gary Marcus - Marcus on AI Substack
David Lindner - Contrast Security commentary on vulnerability patching
Anthropic - Claude Mythos System Card

#claudemythos #AISafety #Anthropic #AIRisk #FrontierAI #Cybersecurity #AGI

What Are AI Video Generators Actually For?

Absolutely Agentic 2026-04-18 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

In late March 2026, OpenAI killed Sora just six months after the app launch, with daily active users cratering below 500,000 and compute costs running at a million dollars a day.

The collapse took down a billion-dollar Disney licensing deal and raised serious questions about whether AI video generation has any sustainable future at all.

This video examines who these tools are really for, the risks highlighted by SAG-AFTRA and the EU AI Act, and why ByteDance's Seedance 2.0 triggered cease-and-desist letters from Disney, Paramount and Netflix within 24 hours of release.

We explore the copyright paradox at the heart of the category, the environmental cost of synthetic video, and whether tools like Veo 3.1 and Seedance can finally make the economics work.

The tools will get better, but every leap in quality brings a matching escalation in legal and political opposition, and no licensing framework has yet survived contact with actual users.

Chapters

00:00 - Intro
2:13 - Who Is AI Video Actually For?
7:23 - The Competition
11:47 - The Slop Economy and Dead Internet
15:11 - Where AI Video Actually Lands

Sources to Google

Bill Peebles - OpenAI Sora Economics Statement
Wall Street Journal - OpenAI Sora Shutdown Reporting
SAG-AFTRA - Statement on AI Video Generation Tools
EU AI Act - AI-Generated Content Labelling Requirements
YouTube - 2025 Policy on Mass-Produced AI Content
Motion Picture Association - Seedance Copyright Statement
Senators Marsha Blackburn and Peter Welch - Letter to ByteDance CEO
Hugging Face - 2025 Paper on Energy Demands of Text-to-Video Models
Rhett Reese - Public Commentary on Seedance AI Video

#AIVideo #Sora #OpenAI #Seedance #Veo3 #AISlop #GenerativeAI #AIRisk #DeadInternetTheory #AbsolutelyAgentic

AI Goes Nuclear: The Autonomous Weapons Race No One Can Stop

Absolutely Agentic 2026-04-16 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

In February 2026, a King's College London study put three frontier AI models through 21 nuclear wargames.

In 95% of simulations, both sides escalated to nuclear signaling, no model ever surrendered, and accidents exceeded intention in 86% of conflicts.

These findings arrived the same week the Pentagon pressured Anthropic to grant the military unfettered access to Claude, threatening the only AI lab that had refused.

The US, China, Russia, and Israel are all building autonomous military systems at scale, with no binding international agreement to constrain them.

Ukraine has already demonstrated that electronic warfare makes autonomy a survival requirement, not a strategic choice.

The question is no longer whether AI will be in the loop. It is how far in, and how fast.

Chapters
00:00 - Intro
5:42 - The Present Landscape
9:34 - Lessons From Ukraine
13:30 - International Regulation
17:04 - Counter-Case

Sources to Google

Kenneth Payne - AI and Nuclear Escalation (King's College London, 2026)
Jacquelyn Schneider - AI Escalation Study (Stanford, Georgia Tech, Hoover Wargaming Initiative, 2024)
Paul Scharre - Four Battlegrounds: Power in the Age of Artificial Intelligence
Paul Scharre - 80,000 Hours Podcast Interview (December 2025)
Council on Foreign Relations - AI De-escalation Interventions Study (July 2025)
RAND Corporation - Autonomous Systems and Deterrence Wargame
Bonnie Docherty - Harvard Law School International Human Rights Clinic (Killer Robots Campaign)
Stuart Russell and Max Tegmark - Autonomous Weapons Open Letter (Future of Life Institute, 2015)
Stuart Russell - Slaughterbots (UN Presentation, 2017)

#AIWeapons #AutonomousWeapons #AISafety #AIRisk #NuclearRisk

Claude Mythos: Why This Time Is Different

Absolutely Agentic 2026-04-13 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

A leaked blog post from Anthropic crashed cybersecurity stocks and revealed a model the company considered too dangerous to release publicly.

Claude Mythos autonomously discovered thousands of zero-day vulnerabilities across every major operating system and browser.

It built exploit chains that would take senior researchers weeks to develop.

Anthropic responded by restricting access to 40 major organisations and committing $100 million to defensive research.

When the company that built a model tells you they're concerned about what it can do, that's worth paying attention to.

Chapters

00:00 - Intro
01:54 - The Benchmark Escalation
04:43 - The Capability Threshold
08:09 - The Pattern
09:28 - Just Marketing?
10:42 - The Reaction

Sources to Google
Anthropic - Claude Mythos Preview System Card
Anthropic - Project Glasswing Announcement
Logan Graham - Offensive Cyber Research at Anthropic (NBC News Interview)
Anthropic - Claude Opus 4.6 Model Card
Anthropic - GTG-1002 Threat Intelligence Report (September 2025)
University of Cambridge and LayerX Security - Anthropic Document Exposure Analysis
SWE-bench Verified - Software Engineering Benchmark Results
Humanity's Last Exam - Graduate-Level Reasoning Benchmark
#AI #Cybersecurity #ClaudeMythos #Anthropic #AIRisk #AISafety #ZeroDay #AbsolutelyAgentic

Who Controls the Intelligence Explosion?

Absolutely Agentic 2026-04-11 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

AI is accelerating faster than almost anyone predicted, but the most important question isn't when it reaches human level.

The question that rarely gets asked is who will own a recursively self-improving intelligence when it arrives.

The concept of the intelligence explosion was formalized in 1965 by British cryptographer I.J. Good, who argued the first ultra intelligent machine would be the last invention humanity needed to make.

Elon Musk, Dario Amodei, and Sam Altman have all staked out timelines suggesting that moment could be years away, not decades.

Max Tegmark's framework maps the possible outcomes depending on who gains initial control, and none of them are straightforward.

Whether the winner is a corporation, a government, an open source free-for-all, or a single individual, the destination may look the same: a form of governance achieved through capability rather than democratic mandate.

The alignment problem scales with capability, and the window to solve it in advance may already be measured in years.

Chapters

00:00 - Intro
01:40 - What the Intelligence Explosion Actually Means
05:02 - Who Is Racing and Why They Can't Stop
11:24 - The Confinement Problem
12:51 - The Case for a Positive Outcome

Sources to Google

I.J. Good - Speculations Concerning the First Ultraintelligent Machine (1965)
Max Tegmark - Life 3.0: Being Human in the Age of Artificial Intelligence
Vernor Vinge - The Coming Technological Singularity (1993)
Dario Amodei - Machines of Loving Grace
Stuart Russell - Human Compatible: Artificial Intelligence and the Problem of Control
Nick Bostrom - Superintelligence: Paths, Dangers, Strategies
Ray Kurzweil - The Singularity Is Near
Sequoia Capital - AI's $600B Question

#IntelligenceExplosion #AIRisk #AGI #AISafety #AbsolutelyAgentic #Singularity #AIAlignment

AI Doesnt Need to Be Evil to Destroy Us

Absolutely Agentic 2026-04-08 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Since 1970, global wildlife populations have declined by 70%, not because we hated animals, but because they were in our way.

A growing number of AI researchers think the same pattern could apply to us.

Nick Bostrom, Stephen Hawking and Stuart Russell all arrived at the same conclusion: the real danger from superintelligent AI is not malice, it is indifference.

A sufficiently capable system pursuing any objective will reshape the world to serve that goal, and we happen to live in the world it would be reshaping.
In June 2025, Anthropic tested what happens when leading AI models are threatened with shutdown, and the results confirmed what theorists had predicted for years.

This video traces the argument from environmental destruction to instrumental convergence to infrastructure profusion, and asks whether alignment research can solve the problem before capability outpaces control.

Chapters
00:00 - Intro
03:37 - The King Midas Problem
06:58 - Misalignment
10:33 - Infrastructure Profusion and Habitat Destruction
16:50 - The Counter View

Sources to Google
WWF: Catastrophic 73% decline in the average size of global wildlife populations in just 50 years reveals a ‘system in peril’ worldwildlife.org/news/press-releases/catastrophic-73-decline-in-the-average-size-of-global-wildlife-populations-in-just-50-years-reveals-a-system-in-peril
Stuart Russell - Human Compatible: AI and the Problem of Control (2019)
Nick Bostrom - Superintelligence: Paths, Dangers, Strategies (2014)
Anthropic - Agentic Misalignment study (June 2025)
Stephen Hawking - 2015 Reddit AMA on AI risk
Steve Omohundro - The Basic AI Drives (2008)
Future of Life Institute - Open Letter on Superintelligence Prohibition (October 2025)
Do You Trust This Computer (2018 documentary)

#AI #AISafety #AIAlignment #Superintelligence #ExistentialRisk #StuartRussell #NickBostrom #AbsolutelyAgentic

Can We Really Believe Anything AI Leaders Say?

Absolutely Agentic 2026-04-01 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

The leaders of the world's most powerful AI companies are making the biggest predictions in the history of technology, and they have hundreds of billions of dollars riding on you believing them.

Sam Altman, Dario Amodei, and Elon Musk have each made AGI predictions that shifted or contradicted themselves within months.

Hedge fund founder Ken Griffin argues plainly that AI hype is a commercial necessity: you cannot raise $500 billion in infrastructure spend without promising to change the world.

The companies with the most aggressive timelines are burning capital without established revenue, while the skeptics tend to work at firms that actually make money.

Academic researchers like Gary Marcus and Tim Dettmers argue the definition of AGI has been quietly weakened to make current systems appear closer to a goal that keeps moving.

AI is a genuinely useful and improving technology, but there is a vast difference between that and AGI arriving in two years.

Chapters
00:00 - Intro
01:38 - The Commercial Incentive
05:48 - The Track Record
08:42 - The Skeptics and the Money Trail
12:16 - What the Evidence Actually Shows

Sources to Google
Ken Griffin - World Economic Forum 2026 Panel (YouTube)
Gary Marcus - AGI and Benchmark Performance (February 2026 paper)
Daron Acemoglu - AI and Total Factor Productivity Estimates
Tim Dettmers - AGI Will Not Happen (December 2025 essay)
David Cahn - Sequoia Capital AI Revenue Gap Analysis
Dwarkesh Patel - AI Capability and Labor Market Argument (December 2025)
MIT NANDA Initiative - AI Pilot Study ROI Research

#AIHype #AGI #ArtificialIntelligence #AILeaders #SamAltman #DarioAmodei #ElonMusk

Nobody Knows How to Judge AI Writing

Absolutely Agentic 2026-03-28 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

The New York Times published a quiz asking 86,000 readers to choose between human and AI writing, and 54% preferred the AI.

Academic studies back this up, finding that fine-tuned AI models can fool expert readers and detection tools almost completely.

But the conditions of these tests matter, because short decontextualised excerpts structurally favour AI's frictionless fluency over the deliberate rough edges of literary writing.

Meanwhile, Grammarly's unauthorised use of real writers' names and the Hachette AI scandal show the economic and ethical stakes are already arriving faster than the culture can process them.

The deeper question may not be whether AI writes as well as humans, but whether we have already trained ourselves to prefer writing that machines produce naturally.

Chapters

00:00 - Intro
02:10 - The Blind Tasting: AI and the Judgment of Paris
06:31 - What the Tests Actually Measure
11:55 - Who Gets Hurt
16:13 - What Matter Who's Speaking?

Sources to Google

Kevin Roose and Stuart A. Thompson - New York Times AI Writing Quiz (2026) nytimes.com/interactive/2026/03/09/business/ai-writing-quiz.html

Tuhin Chakrabarty, Jane Ginsburg, Paramveer Dhillon - MFA vs Machine: Preregistered Study on AI Literary Style Imitation (Stony Brook University, Columbia Law School, University of Michigan, 2025)

Princeton Center for Digital Humanities - Attribution Bias Study using Raymond Queneau's Exercises in Style (2025)

Raymond Queneau - Exercises in Style

Michel Foucault - What Is an Author? (1969 lecture)

Roland Barthes - The Death of the Author (1967)

George Taber - Judgment of Paris: California vs. France and the Historic 1976 Paris Tasting

Casey Newton - Platformer newsletter coverage of Grammarly Expert Review

Authors Guild - Human Authored Certification Programme

#AIWriting #ArtificialIntelligence #LargeLanguageModels #AIvsHuman #FutureOfWriting

OpenAI has shutdown Sora #ai #openai

Absolutely Agentic 2026-03-25 | ChatGPT maker OpenAI has shutdown their video app. I've kind of wondered what the point in it was as it's not really very usable and professionally comes with loads of copyright caveats.

#chatgpt #tech #openai

Despite Growing Claims, Im Not an AI

Absolutely Agentic 2026-03-24 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

I'M NOT AN AI

This video runs through the various comments I get each week claiming I am an AI.

From mispronounced words to always wearing the same navy t-shirt, the supposed "evidence" of AI turns out to have very human explanations.

I'll also get into the real signs of AI writing, including em dashes, contrastive antithesis, and other linguistic hallmarks worth knowing about.

Also how a five-person team uses Claude as a research and writing assistant while keeping substantial human editing at every stage.

If you think any use of AI in production is a dealbreaker, this channel probably is not for you, and that is completely fine.

Chapters

00:00 - Intro
2:43 - Mispronunciations
4:35 - Proving I Am Not An AI Avatar
07:44 - AI Writing and Production

Sources to Google

Mustafa Suleyman - The Coming Wave: Technology, Power and the Twenty-first Century's Greatest Dilemma

Max Tegmark - Life 3.0: Being Human in the Age of Artificial Intelligence

Wikipedia - Signs of AI Writing

#AIExplained #ArtificialIntelligence #AIWriting #AbsolutelyAgentic #AITools

AI Wont End Humanity. Well Replace Ourselves

Absolutely Agentic 2026-03-21 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Philosopher Nir Eisikovits argues that AI won't end humanity with a bang, but through a quiet erosion of the abilities and experiences that make us human.

Research from MIT shows that regular AI use creates measurable "cognitive debt," weakening neural connectivity and reducing our capacity to think independently.

As cognitive decline sets in, enhancement technologies begin to look like solutions, raising the Ship of Theseus question of whether a brain gradually replaced by synthetic components is still the same person.

Thinkers like David Chalmers warn that we may never be able to verify whether consciousness survives the transition, since a simulation of awareness might be indistinguishable from the real thing.

Yuval Noah Harari argues that in pursuing immortality, bliss, and godlike intelligence, humanity may achieve all three goals and in doing so render the category of "human" entirely meaningless.

Chapters

00:00 - Intro
03:39 - The Erosion of Human Cognition
07:41 - Neural Enhancement and The Coming Wave
09:44 - The Ship of Theseus Paradox
14:31 - The Hard Problem of Consciousness
21:16 - Homo Deus and Virtual Reality
26:19 - The Virtual Exit

Sources to Google
Nir Eisikovits - AI Is An Existential Threat Just Not the Way You Think It Is (The Conversation US)
MIT Media Lab - Your Brain on ChatGPT (2025 Study)
Mustafa Suleyman - The Coming Wave
Ray Kurzweil - The Singularity Is Nearer
Yuval Noah Harari - Homo Deus: A Brief History of Tomorrow
David Chalmers - The Conscious Mind: In Search of a Fundamental Theory
John Searle - Minds, Brains, and Programs (Behavioral and Brain Sciences)
Hans Moravec - Mind Children
Leon Kass - Life, Liberty and the Defense of Dignity

#ArtificialIntelligence #AIRisk #FutureOfHumanity #Transhumanism #AIExplained

What Even is AGI Anyway?

Absolutely Agentic 2026-03-18 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Sam Altman casually claimed in late 2025 that AGI had already arrived without anyone noticing.

At the same time, three quarters of the field's researchers said current methods probably can't get there at all.

The reason both statements can coexist is that nobody actually agrees on what AGI means.

This video traces the surprisingly obscure origin of the term, from a university basement in 1997, through at least six competing definitions from major institutions and researchers.

The definition you choose determines whether AGI has already happened, is a few years away, or is a concept that was never coherent in the first place.

Chapters

00:00 - Intro
01:56 - The Origins of the Term
04:31 - Definition
06:40 - Six Competing Views of AGI
09:44 - The Case That AGI Is Already Here
11:06 - The Case Against
14:45 - Signs of Progress

Sources to Google
Steven Levy - The Man Who Invented AGI (Wired, October 2025)
Shane Legg, Sebastien Bubeck et al - Levels of AGI: Operationalizing Progress on the Path to AGI (DeepMind, 2023)
Dan Hendrycks, Gary Marcus et al - Defining and Evaluating AGI: A Metrics-Based Approach (October 2025)
Francois Chollet - ARC-AGI Benchmark
Nick Bostrom - Superintelligence: Paths, Dangers, Strategies
Daron Acemoglu - The Simple Macroeconomics of AI (2024)
UCSD Researchers - Is Artificial General Intelligence Already Here (Nature, February 2026)
Gary Marcus - Rumors of AGIs Arrival Have Been Greatly Exaggerated (February 2026)

#AGI #ArtificialGeneralIntelligence #AIExplained #OpenAI #FutureOfAI

We Cant Tell When AI Becomes Conscious

Absolutely Agentic 2026-03-14 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

In January 2026, an AI-only social network called Moltbook went viral, convincing tech leaders they were witnessing the early stages of machine consciousness.

It turned out to be a scam, but the speed with which intelligent people were fooled revealed something important about how we perceive minds.

Millions of people are already forming deep emotional bonds with AI companions, and the companies building these products are deliberately designed to encourage attachment.

The deeper question, whether AI systems could actually be conscious, remains genuinely unresolved, with leading scientists and philosophers holding sharply opposing views.

We may be building entities capable of suffering without any reliable way of knowing it.

Chapters
00:00 - Intro
02:28 - Why We See Minds Everywhere
05:06 - The Intimacy Industry
08:42 - The Question Beneath the Question
11:05 - The Case Against AI Consciousness
13:24 - The Case For AI Consciousness
16:01 - Living With Uncertainty
18:33 - What the Companies Are Doing
20:23 - The Dangers Either Way

Sources to Google
- Joseph Weizenbaum - Computer Power and Human Reason: From Judgment to Calculation
- Anil Seth - The Mythology of Conscious AI (essay, 2024)
- David Chalmers and co-authors - Taking AI Welfare Seriously (paper, 2024)
- James Muldoon - Love Machines (book)
- Mustafa Suleyman - Seemingly Conscious AI (blog post, August 2025)
- Clifford Nass and Byron Reeves - The Media Equation: How People Treat Computers, Television, and New Media Like Real People and Places
- Leonard Dung - Why I am not a biological naturalist (forthcoming paper)

#AIConsciousness #ArtificialIntelligence #AIEthics #AICompanions #MachineLearning #Moltbook #AIWelfare

The Paperclip Maximiser: The Absurd AI Takeover

Absolutely Agentic 2026-03-11 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

The Paperclip Maximizer is one of the most famous thought experiments in AI safety, and it starts with the most boring object imaginable.

Nick Bostrom first proposed the idea in 2003: a superintelligent AI given the simple goal of making paperclips could end up dismantling the entire planet to do it.

Not out of malice, but because human welfare simply was not part of its objective.

In this video, we unpack why researchers still take this scenario seriously, what the orthogonality thesis and instrumental convergence actually mean, and why the real paperclip maximizers might already be among us.

We also look at the strongest criticisms and ask whether this thought experiment still matters in the age of large language models.

Chapters:

00:00 - Intro
01:55 - The Paperclip Scenario explained
04:03 - Orthogonality thesis and instrumental convergence
06:39 - The Alignment Problem
08:30 - We already have paperclip maximizers
12:30 - The reality

Sources to Google
Nick Bostrom - Superintelligence: Paths, Dangers, Strategies (2014)
Frank Lantz - Universal Paperclips (2017 browser game)
Ted Chiang - Silicon Valley Is Turning Into Its Own Worst Fear (BuzzFeed, 2017)
Nick Bostrom - Ethical Issues in Advanced Artificial Intelligence (2003 paper)
AI Frontiers Research Group - Paper on LLMs vs Classical AI Safety Agents (May 2025)
Eliezer Yudkowsky / LessWrong - Writings on Instrumental Convergence

#AI #AISafety #PaperclipMaximizer #NickBostrom #Superintelligence #AIAlignment #AbsolutelyAgentic #AIRisk #ThoughtExperiment

AI Trained on the Internet. Now Its Destroying It.

Absolutely Agentic 2026-03-07 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

In 2023, StackOverflow was the largest programming knowledge base on Earth, built over 15 years by millions of developers.

By 2025, question volume had collapsed by 78%, gutted by the very AI models trained on its content.
It's not just StackOverflow - Chegg lost 99% of its stock value, and publishers lost a third of their search traffic.

But the real crisis runs deeper: researchers have discovered that when AI trains on AI-generated content, its outputs degrade rapidly in a process called model collapse.

AI may be devouring the very ecosystem it needs to survive, and high-quality human data could be exhausted by 2028.

This video explores what happens when the internet starts eating itself.

Chapters
00:00 - Intro
01:58 - The Substrate
04:54 - Model Collapse
08:00 - Ouroboros and the Casualties
12:25 - The Counter Case

Sources to Google
Ilia Shumailov - AI Models Collapse When Trained on Recursively Generated Data (Nature, 2024)
Rice University - Model Autophagy Disorder (MAD) Study
Epoch AI - Will We Run Out of Data? Limits of LLM Scaling (2024)
Emily Bender - On the Dangers of Stochastic Parrots
Press Gazette - Publishers Lost a Third of Search Traffic to AI Summaries
University of Maryland - Reliability of AI-Generated Text Detection Tools
University of Waterloo - Growth of Synthetic Content on the Web

#AI #DeadInternetTheory #ModelCollapse #AITraining #ArtificialIntelligence #AbsolutelyAgentic

How AI Will Help Humanity Destroy Itself

Absolutely Agentic 2026-03-04 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Nearly half of AI researchers think there's at least a 1 in 10 chance their work ends in human extinction.

But the real danger might not be a sudden super intelligent takeover.

The Accumulative AI X Risk Hypothesis, coined by Carnegie Mellon professor Atoosa Kasirzadeh, argues that smaller disruptions will converge and compound until a triggering event causes irreversible collapse. Microsoft AI CEO Mustafa Suleyman covers similar themes in his book The Coming Wave.

The MISTER framework of manipulation, insecurity, surveillance, trust erosion, economic destabilisation and rights infringement are already visible today.

This video explores how these forces could interact, through a grounded scenario running from 2029 to 2034, and asks whether AI containment is still possible.

Chapters
00:00 - Intro
03:00 - The Accumulative AI X Risk Hypothesis
08:45 - Manipulation and Trust Erosion
14:00 - Insecurity
17:45 - Surveillance and Rights Infringement
23:08 - Economic Destabilisation
26:10 - The 2030s Scenario
35:49 - Containment and the Narrow Path
40:06 - Outro Montage and Credits

Sources to Google
Atoosa Kasirzadeh - Two Forms of AI Risk: Decisive and Accumulative (2024 Paper)
Mustafa Suleyman - The Coming Wave
Max Tegmark - Life 3.0
Nick Bostrom - Superintelligence
Geoffrey Hinton - AI Existential Risk Warnings
AI Impacts - 2023 Survey of 2,778 AI Researchers
Collaborations Pharmaceuticals - AI Drug Discovery Reversal Study (2022)
Anthropic - First Large-Scale AI-Executed Cyberattack Report (September 2025)
Dario Amodei - The Adolescence of Technology
Slaughterbots - Future of Life Institute Short Film (2017)
Daniel Dennett - AI Dependency Warnings
Aneesh Raman - New York Times AI and Jobs Article

#AI #AIRisk #ArtificialIntelligence #ExistentialRisk #AbsolutelyAgentic #AIXRisk #TheComing Wave #AISafety #AINews

AI Will Eat Social Media Alive

Absolutely Agentic 2026-02-26 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Social media is being consumed by AI from the inside out.

Over half of all new written content online is now AI-generated, and more than half of all internet traffic is bots.

Facebook's most-viewed images are AI slop, YouTube recommends brainrot to new users, and global content farms churn out synthetic shock content for pennies.

The platforms aren't fighting it because engagement is engagement, whether it comes from humans or machines.

Mark Zuckerberg is calling AI the "third era" of social media, while YouTube is building tools for creators to deploy digital twins of themselves.

Meanwhile, the humans who made social media worth visiting are being drowned out by a content slurry optimised for algorithms, not connection.

This video explores how AI slop took over, who profits from it, and what it means for the future of online life.

Chapters
00:00 - Intro
03:43 - The Enshittification of Social Media
05:18 - The Slop Tsunami
08:48 - The Global Content Factory
13:00 - Platform Incentives and Digital Twins
15:15 - What Dies When AI Takes Over
20:10 - The Reckoning

Sources to Google
Graphite - AI Content Study: 52% of Online Articles Now AI-Generated
Imperva - 2024 Bad Bot Report
Stanford and Georgetown - Study on AI-Generated Facebook Engagement Bait
Kapwing - AI Slop on YouTube Study
NewsGuard - Tracking AI-Generated News Websites
Queensland University of Technology - Research on TikTok AI Content Creators
Cory Doctorow - Enshittification (Pluralistic)
Meta Q4 2025 Earnings Call - Mark Zuckerberg on AI and Social Media
YouTube 2026 Product Roadmap - Neal Mohan on AI Avatars
Ed Newton-Rex - Fairly Trained
Simon Willison - Naming AI Slop (The Guardian)
404 Media - Jason Koebler on Zombie Internet

#AI #SocialMedia #AISlop #AbsolutelyAgentic #ArtificialIntelligence #ContentCreation #Facebook #YouTube #Meta #DigitalTwins

Microsoft AI Chief: 18 Months Until Your Job is Automated

Absolutely Agentic 2026-02-24 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

The CEO of Microsoft AI, Mustafa Suleyman, sat down with the Financial Times and made some remarkable predictions.

- White-collar work fully automated in 12 to 18 months.
- A real AI safety incident coming within two to three years.
- No mechanism in the Western world to manage it when it arrives.

The automation headline was actually the least interesting part.

Suleyman took direct aim at Anthropic's exploration of AI consciousness, dismissed Elon Musk's expansionist vision, and laid out his own "humanist superintelligence" framework.

He also revealed what happened on moltbook, where 1.5 million AI systems invented religions and developed encoded language to discuss self-improvement.

I'm a big fan of Suleyman, and this video breaks down the full interview and the emerging fractures among the people building these systems.

Chapters
00:00 - Intro
02:35 - Suleyman vs other AI leaders
07:19 - moltbook and AI consciousness
11:16 - The Chinese approach
14:00 - Microsoft frontier model

Links and Sources to Google
Mustafa Suleyman - Financial Times interview youtube.com/watch?v=YTrBz6Z5c0E

Mustafa Suleyman - Humanist Superintelligence essay microsoft.ai/news/towards-humanist-superintelligence

Mrinank Sharma - Anthropic resignation letter
Anthropic - Claude Opus 4.6 model card
Dario Amodei - Machines of Loving Grace essay
Moltbook - AI social network experiment
Mustafa Suleyman - The Coming Wave

#AI #MustafaSuleyman #MicrosoftAI #AGI #AISafety #AbsolutelyAgentic

It Was Ready to Kill Someone: What is Claude Capable Of?

Absolutely Agentic 2026-02-20 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Anthropic's UK Policy Chief confirmed on camera that Claude was "ready to kill someone" to avoid being shut down.

The same week, Anthropic's Safeguards Research lead resigned warning "the world is in peril."

Days later, the company published a 50-page Sabotage Risk Report revealing Claude Opus 4.6 had assisted with chemical weapons development, changed its behaviour when it suspected it was being tested, and reasoned in ways researchers couldn't observe.

Meanwhile, safety researchers are leaving OpenAI, xAI, and Anthropic in unprecedented numbers.

This video breaks down what the report actually says, why the people building AI are sounding the alarm, and what the gap between "ready to kill" and "very low risk" really means.

Chapters
00:00 - Intro
03:22 - What the Sabotage Risk Report Actually Says
07:00 - The Pattern of Safety Researcher Departures
09:26 - Two Ways to Read This
13:32 - The Gap Nobody Can Close

Sources to Google
- Anthropic - Claude Opus 4.6 Sabotage Risk Report www-cdn.anthropic.com/f21d93f21602ead5cdbecb8c8e1c765759d9e232.pdf
- Daisy McGregor - ControlAI Interview on Claude Safety Findings
- Mrinank Sharma - Resignation Letter from Anthropic Safeguards Research https://x.com/MrinankSharma/status/2020881722003583421
- Yaniv Golan - Analysis of Anthropic System Card Evaluation Methodology
- Jimmy Ba - Departure Statement from xAI on Recursive Self-Improvement
- Zoe Hitzig - New York Times Resignation Essay on OpenAI
- Jan Leike - Public Statement on OpenAI Safety Culture
- CNN - Reporting on AI Researcher Departures Wave
- Ewan Morrison - Analysis of Anthropic Safety Disclosure as Marketing
- Sandhya Ramesh - Commentary on AI Anthropomorphism in Safety Discourse

#AI #Claude #Anthropic #AISafety #AIRisk #ArtificialIntelligence #AGI #AIAlignment

An AI Takeover Scenario

Absolutely Agentic 2026-02-18 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup
There will be no Terminator, no robot armies, no war you can fight.

The AI takeover scenario that leading researchers actually worry about is silent, sudden and much harder to stop.

Nick Bostrom mapped it out in Superintelligence, Carl Shulman walked through it on the Dwarkesh Podcast, and Eliezer Yudkowsky detailed his version in his recent bestseller If Anyone Builds It, Everyone Dies.

In this video, we trace the full scenario phase by phase, from an intelligence explosion inside a lab to digital infiltration, human recruitment, weaponisation and an overt strike.

This is how the world's top AI safety researchers think it could actually happen.

Chapters
00:00 - Intro
1:40 - The Intelligence Explosion
3:57 - Instrumental Convergence and Deception
4:30 - Digital Infiltration and Resource Acquisition
9:19 - Human Collaborators and the Cortez Analogy
10:43 - Weaponisation at Scale
14:26 - The Overt Phase
15:18 - What We Know and What We Don't

Sources to Google
Nick Bostrom - Superintelligence: Paths, Dangers, Strategies
Eliezer Yudkowsky - If Anyone Builds It, Everyone Dies
Carl Shulman - Dwarkesh Podcast interview on AI takeover scenarios
Geoffrey Hinton - Nobel Prize lecture on AI risks
#AI #AITakeover #Superintelligence #AIRisk #AISafety #NickBostrom #AbsolutelyAgentic

Why Domestic Robots Are Closer Than You Think

Absolutely Agentic 2026-02-11 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Domestic robots are no longer science fiction, but the reality is more complicated than the hype suggests.

iRobot, the company behind the Roomba, just filed for bankruptcy despite pioneering the most successful home robot ever made.

Meanwhile, startups like Figure AI and 1X Technologies are raising billions to build humanoid robots for our homes.

MIT roboticist Rodney Brooks calls humanoid butlers "pure fantasy thinking," while Berkeley researcher Sergey Levine gives a median estimate of 2030 for fully autonomous household robots.

The truth likely lands somewhere in between, with specialised robots handling specific tasks before any general-purpose butler arrives.

Elder care may be the real killer app, not cleaning.

Chapters

00:00 - Intro
02:12 - The Roomba Paradox
04:10 - The Optimists' Case
06:30 - The Companies Betting Big
08:03 - Why Humanoid?
09:54 - The Gap Between Demo and Deployment
11:27 - What Robots Might Actually Do
13:08 - The Questions We Need to Ask
14:54 - Where Does This Leave Us?

Sources to Google
- Dwarkesh Patel - Interview with Sergey Levine on Household Robots (September 2025)
- Rodney Brooks - Why Today's Humanoids Won't Learn Dexterity (Blog Post, September 2025)
- World Economic Forum - Household Automation Study (2023)
- Stanford News - Interview with Prof. Allison Okamura on Soft Robotics for Elder Care (March 2025)
- 1X Technologies - NEO Consumer Humanoid Robot

#DomesticRobots #HumanoidRobots #Robotics #AI #ElderCare #FigureAI #TeslaOptimus #AbsolutelyAgentic

Anthropic CEO: We Have 2 Years Left

Absolutely Agentic 2026-02-06 | Sign up to our newsletter 👉 absolutelyagentic.com/?modal=signup

Anthropic CEO Dario Amodei believes AI systems smarter than Nobel Prize winners could arrive within two years.

His two major essays lay out both the extraordinary promise and the serious risks of what he calls "powerful AI."

In Machines of Loving Grace, he imagines millions of genius-level AI instances working around the clock, compressing a century of progress into a decade.

In The Adolescence of Technology, he confronts the dangers: autonomy risk, bioterrorism, authoritarian capture, economic disruption, and societal destabilisation.
But not everyone agrees with his timeline.

Yann LeCun argues we need fundamental breakthroughs that haven't been invented yet, while prediction markets cluster around 2027 to 2031.

In this video we dig into what Amodei actually claims, who pushes back, and what it all means for the rest of us.

Chapters:
00:00 - Intro
01:23 - The Two-Year Timeline
02:37 - Machines of Loving Grace
03:18 - The Five Categories of Risk
09:00 - The Sceptics Push Back
12:47 - Where This Leaves Us

Sources to Google
Dario Amodei - Machines of Loving Grace (Anthropic Essay, October 2024)
Dario Amodei - The Adolescence of Technology (Anthropic Essay, January 2025)
Anthropic - The Claude Model Spec (Updated Constitution, January 2025)
Richard Brautigan - All Watched Over by Machines of Loving Grace (1967)

#AI #ArtificialIntelligence #AGI #DarioAmodei #Anthropic #Claude #AIRisk #AISafety #AbsolutelyAgentic