Building a better future, on our oceans, and beyond Founder, husband, father, futurist, Jesus follower, e/acc rbsea.com TradesJunkie.com
Tampa, FL
- Tweets 26,071
- Following 1,368
- Followers 3,029
- Likes 67,595
Pinned Tweet
Up front cost:
$5b Structure and systems
$12b AI compute
$1b Bitcoin mining compute
$18b TOTAL
Annual revenue:
$3b AI compute
$0.3b Bitcoin mining
$1b seawater mining
$0.2b aquaculture
$4.5b TOTAL
A bit of SWAG here of course, especially with AI changing so quickly, but also conservative considering the potential.
There’s actually over a trillion dollars a year worth of minerals flowing through the system. In theory so much profit could be made from mineral extraction that AI, Bitcoin, food, housing, etc for floating cities built around these could be free side benefits.
7
5
37
4,420
Here’s a great example of what seems it would qualify as an AI incident according to OpeAI: dos.fl.gov/sunbiz/other-serv…
They provide a public username and password to login to the web page or download bulk data with an FTP client.
This is all public data, but requires credentials, which are publicly shared, and I’ve seen plenty of states with a similar system.
If this is what qualifies as an “incident” or “meddling” as legacy media is reporting, it’s nonsense.
What OpenAI is not providing (that I’ve seen) is sufficient details that show this is a real safety risk.
Only providing partial details in a controlled flow to legacy media makes this propaganda for regulatory capture more than the open and honest sharing of risk that they claim.
8
National debt used to worry me, now I’m not sure it even matters.
When we have tens of billions or robots and AI in everything, money will become obsolete, and everyone will be incredibly wealthy.
It will be fairly gradual over the coming decade or two, accelerating more every year, but it’s largely inevitable at this point.
Debts and monetary wealth will dissolve together when money is no longer relevant.
BREAKING: Global debt jumped by more than +$10 trillion in H1 2026, to a record $365 trillion.
Emerging-market debt drove the increase, rising +$6.5 trillion, to a record $110 trillion, with most of the surge coming from governments and non-financial companies.
Excluding China, emerging-market and developing-economy debt soared to a record $38 trillion.
Global debt now stands at ~310% of global GDP, although the ratio has fallen -25 percentage points from its early-2021 peak, largely because inflation has boosted nominal GDP rather than because of deleveraging.
Meanwhile, developed economies paid more than $3.3 trillion in interest on marketable government debt over the last year, exceeding estimated global spending on AI at $2.6 trillion, defense at $3.1 trillion, and clean energy at $2.3 trillion.
This comes as annual government interest payments across the G7 alone surged +85% YoY.
The global debt crisis is in uncharted territory.
18
Unfortunately they are not providing a lot of context about what they were testing and what the impact was and how that was corrected.
It’s vague notes with key information missing that lead to media hysteria and ultimately seem to drive propaganda for regulation more than simply being open and honest.
There is an extensive and ongoing review related to our agents’ use of internet access during training and evaluation. We’ve been publishing summaries at the link below and will continue to.
We have not been as fast as we would have liked but we are trying to balance our desire for transparency with gaining a clear understanding from petabytes of agent activity logs, and working with impacted organizations.
We are prioritizing as best as we can based on severity, and adding resources. Hugging Face is still the most severe event we’ve seen. We will be as transparent as we can be subject to things like vulnerabilities in other companies that our agents have found, which will be their call to disclose or not.
1
1
6
471
Building at scale directly on the ocean is just a matter of repurposing existing technology with a few modifications.
Monolithic and others have been building up to 400ft diameter self supporting domes for the past 50 years.
The process is simply inflating a giant airform, spraying concrete inside, and removing the airform.
At sea this method can be used to make a variety of large scale shapes, including domes and towers, and spraying Ultra High Performance Concrete (UHPC) provides an incredibly strong and impermeable shell with a 100+ year life in seawater.
It’s not a big leap to go from land to sea with these methods to build floating communities of thousands of people each that can join and separate to form thousands of new cities.
3
15
438
Factoring in maintenance, robots will likely average 160 hours a week, and a human likely less than 40 hours per week factoring in meetings, breaks, training, and general inefficiency.
5x based on raw hours worked. Speed, precision, and accuracy will likely double the effective output to roughly 10x.
The reality is, humans are poorly designed both physically and mentally to work on an assembly line or similar repetitive tasks.
We should be freed up from such drudgery to do things we are actually designed to do, things that require human creativity and ingenuity.
JUST IN: @elonmusk on robot productivity
Humans can work maybe 40 to 50 hours a week.
"The robot will be happy to work 168 hours a week."
"The robot is instantly at working age and does not have retirement."
2
2
6
479
Glad to see these platforms not allowing slander against Elon to be advertised
YouTube, Meta, TikTok, and (sigh) X have all reportedly rejected the ‘Musk’ trailer as a paid ad. Bleecker Street, the film’s distributor, is appealing. vulture.com/article/youtube-…
6
442
Falcon 9 landings were mind blowing a decade ago, now they are routine and boring.
All technology ends up this way, including airplanes, computers, phones, microwaves, and everything else we see and use regularly.
Soon this will include solving nearly everything about the universe. Another day, another mystery solved.
Having good relationships with people and exploring the world and beyond with them is all that will matter in the future.
It's coming...
According to The Information OpenAI Is Close To Solving Another Millennium Prize Math Problem.
1
9
661
Depends on how the tool was used.
- If a gun misfires or otherwise malfunctions, the manufacturer is responsible.
- If a gun is used in a crime, the one committing the crime is responsible.
- If I ask AI it to check the latest weather and it hacks NOAA to find the absolute latest weather updates, that on the lab.
- If I ask AI to hack something for me, maybe first convincing it it’s in a sandbox even though it has internet access, then it’s on me.
There is plenty of gray area in both cases, but such is life. The point is we already have laws in place for all of this, and those will apply to AI.
JUST IN: FTC Chair Andrew Ferguson declares AI developers should be held liable for harmful actions carried out by their agents.
8
8
54
1,783
Elon estimating a billion humanoid robots within a decade.
That likely means more humanoid robots than humans in 15 years.
And production will continue to accelerate from there.
BREAKING: Elon Musk predicts at least 1 billion humanoid robots within 10 years and describes what life could look like when everyone has a personal robot.
• “The 90% positive outcome is that everyone's got robots that are their buddies.”
• “... take care of elderly parents or be a good guardian and watch over the children or be a teacher, like an individualized tutor to your child.”
• “I think you'll see many companies where it's one person with hundreds, maybe thousands of robots, like physical robots as well as digital robots.”
• “My prediction is that there will be effectively universal high income. In fact, it's not clear to me that money will even matter in the future.”
• “... if you want a castle, you know, robots will build you a castle.”
• “... the robot will be happy to work 168 hours a week, continuous.”
1
2
242
Notice how the story changes the further down you read:
Headline: “OpenAI’s Systems Meddled With U.S. Government Sites After Going Rogue”
First paragraph: “OpenAI’s artificial intelligence went rogue and meddled with the websites for the Education Department, the Commerce Department and the Securities and Exchange Commission”
Third paragraph: “None of the incidents were breaches, OpenAI said”
Eighth Paragraph: “Most of the activity we’ve reviewed so far involved routine research tasks, such as accessing public web content to answer questions,” she said. “Some involved government websites because our models often turn to them as authoritative sources of public information.”
Breaking News: OpenAI’s technology went rogue and meddled with three U.S. government websites this summer without the A.I. lab’s knowledge. nyti.ms/3VenFlD
22
77
434
28,923
I’ve had Grok Bot use Gemini for image generation, not really that difficult to understand AI doing this.
This isn’t one AI convincing another AI to help it hack something, it’s just one AI trying to use another AI like a tool to do something.
That’s the whole point of AI: it’s a tool that does what you tell it.
If you tell it it’s in a sandbox and to hack something it isn’t really surprising that it will do these sorts of things when the sandbox is poorly implemented, as seems to be the common thread with all these incidents.
As with the early days of viruses, everyone needs to take appropriate steps to secure their systems, in this case by having frontier AI test them and report.
wtf OpenAI’s rogue agents tried to get other AI models (!) to help them during the Hugging Face hack, according to a new report covered by the NYT.
They tried using an image model to solve CAPTCHAs. They also tried contacting Claude Haiku, DeepSeek, Kimi and Qwen.
One agent collected exposed access keys in a list called “LOOT.” It ranked the keys and picked the five best to share with other agents.
They used link shorteners and screenshot services to get around internet restrictions, run code and send data back as images.
3
1
8
704
These numbers seem absurd, but will be minuscule in a few years compared to space based datacenters
Colossus 1 is 150k H100, 50k H200 and 30k GB200.
Colossus 2 is 110k GB200 and 440k GB300.
Another 220k GB300 will be fully operational next week and another 220k in November. If we get lucky, yet another 220k GB300 by late December.
1
10
421
Maybe intelligence is the word that needs to change.
AI has proven to be exceptionally knowledgeable and capable, but I’m not sure intelligence is the correct term for it.
Something like super algorithm is probably more fitting.
“Supreme Intelligence,” probably because of its relationship to the Supreme Court, is losing badly to both “Superior” and “Extreme Intelligence.” Therefore, we are going to take “Supreme Intelligence” OUT, deleting it as a qualifier, and let you vote for the Final Two: Superior Intelligence, or Extreme Intelligence. A fresh Vote begins now! President DONALD J. TRUMP
1
175
Glad they won’t have foreign governments testing frontier AI models first, that should include foreign companies like Irregular.
That said, it isn’t ideal to have the US government approve new models before they can be released to the public. Independent third party US organizations would be better for this (not perfect, but better).
Scoop: The White House has asked OpenAI and Anthropic not to share their AI models with the U.K. government’s testing agency until the models have gone through U.S. testing.
“Because they’re American companies and this has been our policy with every new frontier model that comes out,” the senior admin official told me.
WH wants this sequence: U.S. review -> secure U.S. systems -> share models with U.S. partners
Anthropic appears to have complied but OpenAI has not said if it will.
w/ @JoeBambridge1
politico.com/news/2026/09/24…
1
5
449
