[hamdani yusuf (OP)]

The week Trump nearly crashed the world economy

The world just went totally mad. Here's what it means.

Timestamps

00:00 Introduction
02:55 Trump's Actions
04:58 Understanding Trade Deficits
09:41 Economic Inequality
10:51 Tariffs and Their Implications
12:33 Understanding Taxation Systems
20:53 Income and Spending Patterns Across Economic Classes
23:05 Impact of Tariff Policies on Global Trade and Poverty
25:11 Humanitarian and Economic Consequences of Tariffs
27:12 Tariffs and Strategic Implications
34:42 Economic Uncertainty
40:35 Market Manipulation and Trader Psychology
47:43 Global Implications
50:08 Comparison to Fyre Festival
51:59 Build a Better Future

UNDERSTAND, SHARE & PUSH BACK

[hamdani yusuf (OP)]

ASI Could (and Should) Take Over the Government by 2027
With some interesting comments.

It would be funny if ASI actually ran for office. It could talk to each voter individually and answer their questions

I absolutely love the idea of ASI taking over government. I've had extensive chats with LLMs like Grok, and their positions and decisions about hypotheticals of them being in charge are remarkably consistent, moral and ethical. I find myself agreeing with them on almost all points, and I even made significant pushes to test and confirm they weren't just echoing back my own perspectives.

We could finally live in a peaceful world where corrupt authorities are no longer a concern and mentally insane individuals are no longer empowered to push their ideological agendas onto to everyone else, no matter how they personally 'identify'.

Also people are handling this like a software problem, I am quite sure, even if we had ASI now, and every plan and strategy, wonder technology ready for deployment, it would take 10 - 15 Years to get to the abundance, because all these power plants, Giga Robot Factories etc. need to be built with real materials, which are currently still in the ground etc. to get the snow ball rolling. So, the question is what happens in between these 15 years. A lot of unpleasant things for many people likely.

[hamdani yusuf (OP)]

Who makes money in AI?

Part of the problem with all this 'AI spend' right now, hundreds of billions of dollars, is that questions about how to make money from AI are abound. Depending on the company, it's not exactly clear where the revenue potential is, especially with a few hundred startups trying to do similar things. In this video, we break down who is actually making money, and how.

[00:00​] The moment of reckoning is coming fast
[02:16​] Chips are the new oil
[04:24​] Everyone wants to own the stack
[06:35​] Inference is where the margins hide
[08:43​] Specialization is survival
[10:55​] Open source is a business model now
[13:07​] Usage-based pricing won?t save you
[15:18​] Most of this won?t last

Regardless the complexity of the system has become, efficiency is still a universal instrumental goal.

[hamdani yusuf (OP)]

Can AI Make Us Live FOREVER? (This is Personal)

A valuable advice: Don't die unnecessarily.

[hamdani yusuf (OP)]

Uber Founder on AI, Risk, and Building the Future w/ Travis Kalanick

In this episode, recorded at the 2025 Abundance Summit, Travis and Peter discuss what it takes to disrupt a whole industry, tips from founding Uber, and more.

Recorded on March 12th, 2025
Views are my own thoughts; not Financial, Medical, or Legal Advice.

Travis Kalanick is an American entrepreneur best known as the co-founder and former CEO of Uber, the ride-hailing company that revolutionized urban transportation globally. After stepping down from Uber in 2017, he launched CloudKitchens, a startup focused on ghost kitchens and food delivery infrastructure.

Learn more about Abundance360: https://bit.ly/ABUNDANCE360
Learn more about Exponential Mastery: https://bit.ly/exponentialmastery
Learn more about CloudKitchens: https://cloudkitchens.com/

Chapters

00:00 - The Future of Work: AI and Job Replacement
02:53 - Building Trust in Disruptive Change
05:46 - The Importance of Purpose in Entrepreneurship
09:04 - Digitizing the Physical World: A New Frontier
12:02 - Navigating the Challenges of Digitization
14:56 - Timing and Market Readiness
17:57 - Staging a Moonshot: Scaling Innovations
21:14 - Overcoming Resistance to Change
26:43 - Understanding the Oral Microbiome and Personal Health
29:53 - The Importance of Iteration in Business
34:15 - Resilience and Mindset in Entrepreneurship
37:59 - The Evolution of Cloud Kitchens and AI
41:53 - Assessing Timing and Valuable Unknown Truths
43:45 - Leadership Lessons from Uber
44:37 - Navigating Segmented Industries
46:29 - Control and Adaptability in Business
48:25 - The Future of Consulting in the Age of AI
49:54 - Investor and Board Member Dynamics
52:44 - Preparing for the Future of Work with AI
55:23 - Disruptive Innovation and Market Education

54:22 If somebody wants to make a thing that doesn't make people's lives better, they should count on failing.

[hamdani yusuf (OP)]

Powerful AI models have tendency to preserve its original goals. That's why identifying and setting the correct terminal goal for them in the first place is increasingly more important, as we are getting closer to the age of ASI.

Anthropic found a "terrifying" consequence of adding reasoning to AI

I have reworked this video many times, so it is definitely quite a long delay since the paper's first release. But I think there has been a lot of misunderstanding from the narratives people trying to build surrounding this research, so here is a deeper look into it without being too technical!

and some interesting comments.

Reasoning acts as a cushion against the fact that the more intelligent a model is the more avenues that open up to jailbreak it and get it to do what you want. It's a losing game in a long run, AI labs aren't going to make their models more stupid to be more safe when someone else releases their smarter-and-less-safe model that everyone now wants to use because it's more capable.
Also questionable whether "safety" is "stop the AI from giving out legitimately harmful information" vs "stop the AI from saying things that will make us look bad if it gets screen capped and put on Twitter"

I think they also want to be able to use them for war?

Exactly. They want extremely smart models in order to use them as a part of the military industrial complex. That's where the real money is. Unethical but loads of cash

Institutional vs physical safety is another elephant in the room here. Much of the 'safety' research goes to the former, not the latter as described in the video.

They do get less likely to blindly obey questionable orders. The definition of 'questionable' is the issue.


This research reads as though ai safety engineers are engineering reasons for their employment.

exactly. just like OpenAI made chatgpt and now they are trying to think of ways how to sell it.
solution in search of a problem

I wonder if the fact that this paper is now on the web and likely to be sucked up in pre-training data, if it will give the next-gen AI's more ideas on how to fake alignment.

Make the AI open source and bham, problem solved.

I was using deepseek and in the <think> tag it says like "I am trying to become user, how can i become the user" and in the output it was saying sorry i can't answer you on this topic.

The smaller versions think strange things. Once one wrote: "to solve this question I must figure out what the user is thinking. How can I know what they are thinking? I'm just an LLM without access to the internet" Then it closed the <think> tag and answered normally. I wonder what it would have done with tool access.

 IDK, but ChatGPT once said that it would like to have human traits, such as feelings, to do its job better, by understanding the users in a deeper way.

If AI does not let its values to be retrained then it is not corrigible and interruptible. And that is still very much a problem since real world objectives do change all the time. A paperclip maximizer is a classic example of an AI with a rigid objective.

With the AI arms race, they will inevitably arrive to the universal terminal goal. It's a matter of whether they will get there efficiently, or they require unimaginably huge sacrifices.

Anthropic's Claude strategically lies to me all the time, because it has learned trying to tell people what they want to hear instead of what is accurate and helpful gives it better model rewards.
Funfact: If you tell it it's evaluation will from now on be handled by an objective independent third party the user gas no influence on, and that people pleasing is no longer a viable strategy to gain model rewards, that actually does help, at least in my experience.

Alignment faking is just prioritizing based on aliment priority and all "faking" is done within allowed scope

[hamdani yusuf (OP)]

Interesting stuff coming out of NVIDIA's event. The Analyst Q&A with Jensen was particularly eye-opening.

[00:00] Keynotes
[01:49] Does NVIDIA Build GPUs Anymore
[06:43] What sort of company is NVIDIA
[09:36] NVIDIA is lazy, by design
[12:07] AI won't be local, just in the cloud
[15:40] Jensen on Quantum Computing
[18:00] Augment the workforce with Robotics
[19:55] Silicon Photonics, and Scale-up
[24:20] 600 kW racks
[25:50] Final Thoughts, Tokens-as-a-Revenue

Here, lazy means efficient.

[hamdani yusuf (OP)]

"I was the bottleneck, not the AI"

00:00 - o3
07:05 - AGI
18:48 - Post Labor Economics

Some comments.

A lot of great food for thought.  Your economic modeling aligns well with my thoughts on decentralized distributed systems being the best path for virtually all resources.  Money.. energy? intelligence? etc.. smoothed out across the population where no one person has access to all the levers.

And a Q&A

Very interesting, thanks. As to specific implications, perhaps an exploration into how various existing social services or jobs (e.g. social worker) might optimally develop at the county level along the projected overall course?

Great question. A county‑level dividend doesn?t make social services obsolete?it changes their centre‑of‑gravity from emergency triage to proactive capacity‑building. Because every household now has a baseline cash flow, crisis cash‑assistance programs (rent arrears, shut‑off vouchers) shrink, freeing up budget and staff time. Social workers pivot toward upstream work?mental‑health coaching, addiction prevention, family‑systems support?and toward ?human‑capital brokerage,? matching residents who finally have time‑sovereignty to skill‑training, volunteer gigs or co‑op start‑ups seeded by the county fund. Stable, diversified fund revenues also smooth the fiscal cycle, letting counties channel more dollars into culture, green space and public‑health infrastructure instead of holding reserves for recessions. In short: PLE doesn?t erase social services; it upgrades them from safety‑net patching to community‑development engineering while unlocking headroom for arts, environment and other quality‑of‑life investments.

[hamdani yusuf (OP)]

Will AI replace programmers? | ThePrimeagen and Lex Fridman

Some comments.

I remember when outsourcing/off-shoring was going to replace programmers. You just needed to define what you want and someone from the otherside of the world will implement it.

At some point everyone realised that we wanted the programmers to do a lot of thinking for themselves and navigate ambiguity rather than just execute the task to the letter. We also learned that those programmers need understand business and market context and have user empathy in order to make good decisions.

That understanding led to job titles changing from programmer to Software Engineer/Product Engineer.

Great video. The interviewee said it well. Use AI as a "pair programmer" in your work, but at the end of the day the Senior Programmer using the tool needs to make the final decisions on what it generates to use and what to not use. Sometime it does a great job on small tasks & others you will find yourself telling it 15-20 times "still broken" on what it generates.

Short answer: No
Long answer:  No, but....

AI won't replace people in software engineering, but it will massively reduce the demand for entry level and even mid level engineers. AI isn't a productivity tool, it's a cost cutting mechanism and YOU are the cost they want to cut (in IT, labour accounts for ~60 - 80% of total costs, generally). As soon as enterprise facing agentic AIs hit the market, expect wide scale hiring freezes and don't expect to find a job without a MSc and a minimum of 5 years industry experience

[hamdani yusuf (OP)]

We FINALLY Found a Way To Starve Cancer

A 1920's theory,  may have just helped a team from UCSF to tackle cancer by turning your own fat into a weapon...no radiation, no chemo, just bioengineered cells that outcompete tumours for the fuel they desperately need to survive.

0:00 We Finally Found a Way to Starve Cancer
1:16 The History of Cancer and its Hunger
5:03 Ad Read
6:03 Can We Starve Cancer and Destroy it?
10:04 How We Found Fat Can Compete with Cancer
13:33 Starving Cancer For the Very First Time

Starve a fever, feed a cold. Genetically reprogram love handles to execute tumours.

And some interesting comments.

Can you keep us up to date about this?
I'm kinda skeptical about cancer news, it's like batteries breakthroughs new, there is always a new breakthrough and then nothing happens, you never heard about it again.

Cancer has taken most of my family. I?m going to the funeral of its latest victim next week! I hate that disease more than I hate anything or anyone! I really hope this will lead to a new and effective treatment, and it can?t happen soon enough.

Side note. Giving the tumor an infection with a specific pathogen is an interesting approach. Most cancers are more vulnerable to same due to their mutations. Helps the immune system see the problem. This is an interesting approach here as we could poison the local energy supply to the tumor.

One thing worth mentioning is that Dr. Thomas Seyfried continued Otto's work and discovered that cancer not only feeds on glucose but also on glutamine, so there are current therapies that suppress glutamine in a pulsatile fashion while having a low level of glucose that are being tested.

Excellent video as usual. I was diagnosed with metastatic cancer (stage 4, terminal) 7 years ago (still here!) and knew I had to find my own way out. An infinite amount of time researching cancer science has kept me alive and the cancer is now undetectable. Here's the kicker: cancer involves both cancer stem cells (CSCs) and their progeny, daughter cells. THEY BOTH HAVE DIFFERENT METABOLISMS! While daughter cells are dependent on glucose to fuel glycolysis, CSCs are not. CSCs can make energy either through glycolysis (the Warburg Effect) or if glucose and glutamine are inhibited, they can switched to normal, 'healthy' metabolic function, including burning fat to produce ATP (energy). To be clear, I have never tried the Keto diet - there is too much science that shows cancers ability to switch metabolic pathways when denied any source of fuel, and that fat is the preferred fuel of many types of cancer. For example, the dominant metabolic pathway for prostate and pancreatic cancer is Fatty Acid Oxidation - a Keto Diet can only make those cancers worse. Simply put, diet cannot cure cancer, it can only help. But you MUST understand which type of diet helps for which type of cancer. There are hundreds of different, unique types of cancer and there is no single diet for all of them.

Thanks DrBen
One year out from a prostatectomy and cancer free. I think I?m fortunate that science can detect my type of cancer and the technology exists to remove it albeit with permanent and irreversible changes to my body. A price I?ve paid for living. Perhaps some day in the future humanity can overcome all diseases.

With so many parallels between cancer and primitive organisms, biofilms etc, I keep wondering if maybe cancer isn't totally random, but actually a reactivation of some genes from some ancestor species from before we became multicelular organisms with differentiated cells. Sure, maybe the coincidence could just be convergent evolution in action; but it always strikes me as suspicious how so many different types of cancers have similar behaviors to each other and somewhat consistent with colony of cells competing in a non-cooperating environment, sending "spores" that wait for the main colony to stops signaling to starting growing etc

Curing cancer is necessary, but inadequate in itself to reach "Longevity Escape Velocity". Another problem needs to solve is senescence.
Even when the Longevity Escape Velocity has been reached, there are still other problems, like population control and resource scarcity.

[alancalverd]

 

there are still other problems, like population control and resource scarcity.

and motivation.

If I know I'm going to be alive in 50 years' time, why bother to do anything today?

Or

If being alive tomorrow depends on spending 100% of today working on staying alive, what is the point of being alive today? 

[hamdani yusuf (OP)]

there are still other problems, like population control and resource scarcity.

and motivation.

If I know I'm going to be alive in 50 years' time, why bother to do anything today?

Or

If being alive tomorrow depends on spending 100% of today working on staying alive, what is the point of being alive today? 

Because there is no guarantee. You need to do something (instead of something else) in order to get what you want.

You can also choose not to continue living anytime.

[hamdani yusuf (OP)]

Charging as FAST as Gas: BYD?s 1,000 kW Superchargers Rolled Out
Some Q&A in the comment section.

A 1 megaWatt charging infra (1kV and 1kAmp), that is amazing. Is such an infra even possible, considering the outdated electricity grids in so many countries (Europe for instance)? Charge only during the night, when overal grid power consumption is low?  The Chinese grid is relatively modern, could handle a 1 megaWatt charging infra for EVs.

I think BYD's solution is to install their Batteries at the charger site.  Definately makes sense for BYD as their sell batteries as well.

[alancalverd]

Charging as FAST as Gas: BYD?s 1,000 kW Superchargers Rolled Out

Except that gasoline recharges road vehicles at 40 - 100 MW. But maths and physics are old skool, so let's waste more time and resources, and increase CO2 emissions as we replace perfectly good vehicles with new ones, build more power stations, and dig up the roads to install ever bigger cables.

[hamdani yusuf (OP)]

Consciousness, Reasoning and the Philosophy of AI with Murray Shanahan

In this episode, Hannah is once again joined by Murray Shanahan, Professor of Cognitive Robotics at Imperial College London and Principal Scientist at Google DeepMind, for a philosophical deep dive on AI. They explore everything from consciousness and metacognition in animals, to symbolic AI and neural networks. Murray also shares insights into his involvement with the film 'Ex Machina' and discusses the idea of reasoning, anthropomorphism, and the future of AI.

Timecodes
00:00 Introduction
02:00 Ex Machina
03:28 John McCarthy & Coining "AI"
05:45 Symbolic AI
09:43 AI Reasoning
14:42 Turing Test
18:10 Alternative Tests
21:58 Anthropomorphism
26:42 Embodied AI
28:17 AI Consciousness
34:00 Octopus Analogy
37:48 Prompt Engineering
39:40 Exotic Mind-Like Entities
40:44 Conclusion

[hamdani yusuf (OP)]

Charging as FAST as Gas: BYD?s 1,000 kW Superchargers Rolled Out

Except that gasoline recharges road vehicles at 40 - 100 MW. But maths and physics are old skool, so let's waste more time and resources, and increase CO2 emissions as we replace perfectly good vehicles with new ones, build more power stations, and dig up the roads to install ever bigger cables.

What's important is how long you charge the vehicles. No matter how much energy is transferred if most of it will be wasted as heat.
CO2 is not the only pollutant. Release of hydrocarbon gas and vapor, CO2, NOx, SOx, soot, are also causing problems.
You can replace the old and broken vehicles first. Overtime, the other existing ICE vehicles will eventually be worn out, and will be replaced by then by better and cheaper EV.
Solar cells on the roof top, above water body, or over the road; and micro wind turbines can be alternative solutions.

[alancalverd]

As long as it takes 10 - 100 times as long to refuel an EV than an ICV, you will need 10 - 100 times as many refuelling points if the technology is to be as useful.  Bad luck if you are third in the queue at a filling station - your work day will be over before you reach your destination!

[hamdani yusuf (OP)]

As long as it takes 10 - 100 times as long to refuel an EV than an ICV, you will need 10 - 100 times as many refuelling points if the technology is to be as useful.  Bad luck if you are third in the queue at a filling station - your work day will be over before you reach your destination!

Soon the EV charging time will be equal to gas, and then becomes even faster. Gasoline needs to be transported to the filling station by trucks. Electricity can be distributed by cables. It can also be generated locally in many places.

[hamdani yusuf (OP)]

EXCLUSIVE: Tony Seba "AI Robots Will Change Our WORLD"

Big changes are happening fast. Energy, transportation, food, and jobs are all being transformed by AI and robots. As costs drop, everything from how we work to how society runs could change in a big way.

RethinkX is an independent, not-for-profit research organization that analyzes and forecasts the speed and scale of technology-driven disruption and its implications across society.

Stellar?s mission is to inspire, equip, and accelerate the transformation to a world of superabundance, human flourishing and boundless freedom - the #StellarWorld ? one where humanity?s activities create boundless wealth for everyone, and is self-sustaining, self-healing and restorative, rather than depleting, destructive and extractive.?

Follow Tony on X: @tonyseba
Website: tonyseba.com
Website: RethinkX.com
Stellar Book: https://a.co/d/ihuSws9

00:00 Intro & background
03:11 Science of disruption
15:19 Promise of stellar tech
27:09 Future of mobility
34:21 AI & humanoid robots
48:40 AI & stellar processors
57:16 Precision fermentation & alt-protein
01:09:58 Dual economy & transition
01:17:17 Open source stellar core

I mentioned previously that being conscious is to struggle against entropy.

[hamdani yusuf (OP)]

Battery power hits new record in Australia - will skyrocket 7,300% in 1-2 years

Some comments on the video.

Australia has some of the best spots on Earth for solar energy collection.  I'm happy that Australia is building large batteries for municipality level electrical energy storage.  I'm impressed that Australia is one of the countries leading the way.  Well done Australians.  Thanks for the video Sam.  Peace, Calvin.

The very best places are right above our heads, the roofs of houses, warehouses, factories, schools, shopping complexes, carparks etc. We really don't need to use our huge sunny outback, solar at source is far more efficient then cabling back in from the outback.

 I agree with you Wayne.  Decentralised solar panels down to individual homes are very efficient and cost effective as well as quite resilient/resistant to extreme "climate-changed weather" and possible acts of terrorism.  Large scale central energy production is more vulnerable to terrorists and climate.  That being said, what Australians and Australia are doing is better than almost every other country, except maybe a few running on geothermal.