Silicon Investor - Artificial Intelligence, Robotics, Chat bots - ChatGPT

[Frank Sully] Kraken supplies batteries and SS Sonar to Anduril. Anduril ramps up underwater v...

Frank Sully — 8/30/2026 9:03:28 PM

Kraken supplies batteries and SS Sonar to Anduril. Anduril ramps up underwater vehicles production.

defensedaily.com

[Elroy] Q3 and Q4 will be crucial indications of how the integration of Covelya and Krak...

Elroy — 8/27/2026 1:46:10 PM

Q3 and Q4 will be crucial indications of how the integration of Covelya and Kraken Legacy is working out.

Doesn't it depend on the length start to finish of the sales cycle?

Are these undersea robots sold in a month, or is there a multi-month evaluation, followed by a massive purchase order?

[Elroy] Learn something new every day......

Elroy — 8/27/2026 1:44:55 PM

[Frank Sully] Kidl, Indeed Q3 and Q4 will be crucial indications of how the integration of Cov...

Frank Sully — 8/27/2026 1:25:08 PM

Kidl, Indeed Q3 and Q4 will be crucial indications of how the integration of Covelya and Kraken Legacy is working out. Also we'll see how growth of Anduril business is going.

[Frank Sully] PNG is the Caadian symbol for Kraken Robotics.

Frank Sully — 8/27/2026 1:17:10 PM

[Elroy] What's PNG?

Elroy — 8/27/2026 12:47:48 PM

[kidl] PNG ... The next couple of quarters will be the ones to watch.

kidl — 8/27/2026 12:47:09 PM

[Elroy] Kraken Robotics Reports Q2 2026 Financial Results finance.yahoo.com Since its ...

Elroy — 8/27/2026 12:27:00 PM

Kraken Robotics Reports Q2 2026 Financial Results

finance.yahoo.com

Since its news release on July 2, 2026, the Company has secured additional product orders totaling over $27 million across a number of technology products including navigation and positioning systems, SAS, monitoring solutions, including its intruder detection system and seabed pressure sensors, and platform sales of its ScanFish remotely operated towed vehicle (ROTV). The company has also delivered a record number of Vigilant Forward Looking Sonars (FLS), which can be integrated on UUVs and USVs for obstacle avoidance. Announced orders in 2026 now total approximately $355 million on a combined basis for Kraken and Covelya Group.

[Frank Sully] Beth Kindig on NVIDIA 2Q io-fund.com

Frank Sully — 8/27/2026 11:23:16 AM

[Frank Sully] NVIDIA Knocks The Cover Off The Ball Again - Up 5% AH seekingalpha.com

Frank Sully — 8/27/2026 2:49:26 AM

[A.J. Mullen] I agree with your last sentence. It's not impressive that a man-made machine can...

A.J. Mullen — 8/26/2026 10:01:24 AM

I agree with your last sentence. It's not impressive that a man-made machine can travel 100m faster than any human. Stephenson's Rocket could do that in 1829, en.wikipedia.org. It could maintain that speed over miles. It's a gimmick to claim that a robot has broken any human record. Better automation avoids anthropomorphism.

[Savant] World humanoid robot games show runners breaking records, breaking apart, and bu...

Savant — 8/26/2026 7:18:20 AM

World humanoid robot games show runners breaking records, breaking apart, and bursting into flames

It's a learning experience.

arstechnica.com

But not all robotics researchers are convinced that humanoid robots are necessary, especially when the most useful robots currently doing productive work usually come in other shapes and sizes.

[Frank Sully] A.J., It is probably true that humanoid robotics profitability is years off. Goo...

Frank Sully — 8/25/2026 6:04:45 PM

A.J., It is probably true that humanoid robotics profitability is years off. Google's AI lab DeepMind's Gemini Robotics is state of the art AI and shows clearly how long we have to go.

youtu.be

Google has been funding robotics research for a decade, going back to Everyday Robots. Google's massive Search profits have subsidized Moon Shots like these. Decades of research into AI are being monetized as Gemini enhances Search, Cloud, YouTube and DeepMind. Google's autonomous auto subsidiary Waymo robotaxis are expanding to more and more cities and will reach profitability in several years.

Fortunately for my favorite AI & Robotics play NVIDIA, making up 50% of my million dollar nest egg, being a vendor of picks and shovels (chips and networks) and maps (CUDA software libraries), it makes substantial profits off its robotics hardware and software.

seekingalpha.com

[A.J. Mullen] This article in The Economist predicts profits are far-off for humanoid robots. ...

A.J. Mullen — 8/25/2026 10:05:31 AM

This article in The Economist predicts profits are far-off for humanoid robots. economist.com Training is identified as a high barrier to human-like adaptability. Massive amounts of data are needed, which are sparse and expensive.

[Savant] Video: Humanoid robot rides 660-pound capacity robot horse in Beijing demonstrat...

Savant — 8/24/2026 9:18:59 PM

Video: Humanoid robot rides 660-pound capacity robot horse in Beijing demonstration The Qiji X1 combines mechanical legs, a saddle and handlebars to turn a quadruped robot into a rideable machine.

Humanoid robot mounts 660-pound capacity robot horse in Beijing expo

[Frank Sully] NVIDIA Vera Rubin NVL72 Delivers 30x Efficiency Boost for AI blockchain.news

Frank Sully — 8/24/2026 1:14:33 PM

[Frank Sully] NVIDIA - Groq Low Latency Inference Chips In Full Production; Will Revolutionize...

Frank Sully — 8/24/2026 11:20:48 AM

NVIDIA - Groq Low Latency Inference Chips In Full Production; Will Revolutionize Agentics

blogs.nvidia.com

[Savant] BEIJING -- Chinese humanoid robots broke marks set by humans, including beating ...

Savant — 8/22/2026 2:58:43 PM

BEIJING -- Chinese humanoid robots broke marks set by humans, including beating Usain Bolt's 100-meter sprint world record, on the opening day of the Olympics-like World Humanoid Robot Games in Beijing on Saturday.

More than 2,000 humanoid robots are participating in the event, the organizer said

Humanoid robots surpass human records in 100m, high jump

Humanoid robots surpass human records in 100m, high jump

[Frank Sully] More on NVIDIA Robotics Mid-July brought the "Cosmos Coalition," a group includ...

Frank Sully — 8/18/2026 5:44:48 PM

More on NVIDIA Robotics

Mid-July brought the "Cosmos Coalition," a group including Fujitsu, FANUC, Yaskawa Electric, Kawasaki Heavy Industries, Hitachi, NEC, SoftBank, Sony and Kubota. These industrial heavyweights will build on Nvidia's Cosmos, Isaac, Metropolis and Jetson platforms to advance physical AI applications in robotics and factory automation.

CFO Coletter Kress, FY2026Q4

<<< Physical AI is here having already contributed north of $6 billion in NVIDIA revenue in fiscal year 2026. Robotaxi rides are growing exponentially with commercial fleets from Waymo, Tesla, Uber, WeRide and Zoox, and many others are expected to scale from thousands of vehicles in 2025 to millions over the next decade, creating a market poised to generate hundreds of billions of dollars of revenue. This expansion will demand orders of magnitude more compute with every major OEM and service provider developing on NVIDIA's platform.

We continue to advance robotics development. With the new NVIDIA Cosmos and Isaac Group, open models, frameworks and NVIDIA's powered robots and autonomous machines for leading companies, including Boston Dynamics, Caterpillar, Franka Robotics, LG Electronics and NEURA Robotics. To accelerate industrial physical AI adoption, we also announced new expanding partnerships with Dassault Systemes, Siemens and Synopsys to bring NVIDIA AI infrastructure Omniverse digital twins, World Models and CUDA-X libraries to millions of researchers, designers and engineers building the world's industries. >>>

CFO Colette Kress FY 2027Q1

<<< Our physical AI continues to gain momentum, exceeding $9 billion in revenue over the last 12 months. Our partnership with Uber will power the robotaxi fleet across nearly 30 cities and 4 continents by 2028. And in robotics, leading companies across a range of industrial, surgical and humanoid applications are building on NVIDIA's technology to develop and deploy at scale. We remain front-footed in securing sufficient supply to support our customers' growth. >>>

[Frank Sully] Nvidia and LG Electronics plan to speed up commercialization of robotics Nvidia...

Frank Sully — 8/18/2026 12:47:49 PM

Nvidia and LG Electronics plan to speed up commercialization of robotics

Nvidia (NVDA) and the Seoul-based LG Electronics signed a memorandum of understanding last week to speed up the commercialization of robotics.

LG will utilize Nvidia's robotics stack, which includes the Omniverse libraries, Cosmos open world models and the Isaac robotics development platform. LG is also building a Data Factory where robots collect data through physical learning. It is expected to house hundreds of robots by the end of the year.

The training is expected to produce hordes of data to continuously improve the capabilities of the robots. The data will advance LG's Robot Foundation Model, which functions as the brain of the machines. The robots are mainly focused on household chores and manufacturing work. LG also established a Robotics Business Center last month for its commercialization efforts.

"Through the synergy built on 'One LG'—bringing together core capabilities across the group—and strategic collaboration with global partners, we will secure our competitiveness in physical AI and become a comprehensive robotics solutions provider," said LG CEO Lyu Jae-cheol.

[A.J. Mullen] I asked Claude if it had access to the Morgan Stanley report. It didn't have dir...

A.J. Mullen — 8/17/2026 8:47:13 PM

I asked Claude if it had access to the Morgan Stanley report. It didn't have direct access but it did have detailed press coverage. My input is in bold - I love AI's tolerance of typos! The italics are Claude's response:

OK. I don't undertand how they can be confident of their estimate of the margins for API services given that Western companies are switching to Chinese service providers becvause they are cheaper and teh Financial Times recently stated Open Ai and Anthropic were in a price-war having substantially reduced their prices for their 2nd tier product, and I have never paid a penny for any AI model.

Searched the web

Searched the web

Your skepticism is well-founded — the reporting confirms exactly the tension you're describing, and it's worth laying out where it actually bites the Morgan Stanley model versus where it doesn't.

What's confirmed:

OpenAI has cut pricing on GPT-5.6 Luna substantially, and Anthropic has positioned Claude Opus 5 at roughly half the price of its higher-end Fable 5 model, while Google's Gemini 3.7 Flash launched August 13 at introductory rates well below prior norms.
More specifically, OpenAI cut GPT-5.6 Luna by 80% and Terra by 20% on July 30, 2026, attributing it to roughly 20% efficiency gains in serving costs.
Chinese open-weight models grew from under 2% of token consumption on OpenRouter in late 2024 to more than 50% by June 2026, and DeepSeek V4 Pro prices output at roughly 1/57th the cost of Fable 5 — that's the "switching to Chinese providers" dynamic you're describing, and it's real and large.
There's a live analytical question about whether the price cuts are funded by genuine efficiency gains or by outside capital, given OpenAI is losing billions while cutting prices and carries roughly $665 billion in cloud commitments through 2030.

Where this actually lands on Morgan Stanley's numbers: their $1.75/million token baseline for the "owned infrastructure API" bucket (the 46% ROIC case) is already toward the low end of current market pricing for frontier-tier models, not the high end — their own sensitivity table shows the business loses money below $1/million and only reaches ~32% margin at $2.50/million, so a further slide toward DeepSeek-style pricing ($0.03–0.05/million for some Chinese models) would blow straight through their loss threshold if premium labs were forced to match it dollar-for-dollar. Their framework doesn't really contemplate that scenario — it's modeling continued pricing power for proprietary/frontier-tier tokens, not a race to Chinese-open-weight price levels.

The report's actual defense (per the cuts reported by FT) would probably be that this is tiered, not uniform, price discrimination: Luna/entry-tier models get commoditized fast (which MS's framework doesn't really price for), while flagship tiers like Sol or Opus retain pricing power because enterprise buyers aren't purchasing purely on cheapest token rate — reliability, tool use, context handling, and integration lock-in matter more than raw $/token. That's a real distinction but it's also exactly the assumption that's untested at scale, which is the core of your objection. Milk Road

On you personally never having paid a penny: that's not actually in tension with the ROIC model — free-tier consumer usage (like what you and I are doing right now) is typically loss-leader/subsidized by enterprise API revenue, not counted as the revenue base MS is modeling. The report's ROIC estimates are built on enterprise/API token pricing, not consumer subscription or free-tier economics, so your free usage doesn't disprove their numbers — but it does underline that a huge share of actual usage volume generates zero revenue, which is a real gap between "tokens served" and "tokens monetized" that the report's clean per-GW revenue assumptions gloss over.

Bottom line on your challenge: the report's own sensitivity analysis already shows its central estimate sits close to the point of unprofitability, and it was published before the July 30 cuts and the FT price-war piece — so if you're asking "does Morgan Stanley's confidence hold up against what's happened since late July," the honest answer is: not obviously, and the report itself gives you the numbers to see why a further slide in flagship-tier pricing (which is exactly what's been reported) would compress or eliminate the margin it's projecting.

[A.J. Mullen] That's a useful contribution to the debate. Citron's analysis is a little breath...

A.J. Mullen — 8/17/2026 1:53:11 PM

That's a useful contribution to the debate. Citron's analysis is a little breathless, but the fact that Nvidia is stimulating or smoothing the purchase of its products is a concern for me. It's end-users who provide the ultimate justfication. I use AI daily but have yet to pay a penny for the privilege. Western companies, who are paying, are said to be switching to Chinese models because they are much cheaper. The price to revenue of the Chinese companies is much higher those of the US, so in a sense the access to the Chinese models is cheaper because they are being more heavily subsidized by investors. At some stage, users of AI have to pay the full cost. That's no where near true yet. Open AI and Anthropic are in a price-war for users, ft.com.

All this investment will reduce the cost to the end-user. While each generation of chip is more expensive as well as more powerful, crucially, the power is increasing faster than cost. At some time, providers of AI will be able to make money on selling their service. Whether those profits will justify their current valuations is an open question. I hold NVDA. It's earnings are very real and it it has a big order backlog, but if the music stops, many of the orders will be cancelled and some of future earnings will go to compensate lenders.

[Frank Sully] Gib, The wave of continuously expanding AI capital expenditure, which has spark...

Frank Sully — 8/15/2026 2:02:02 AM

Gib,

The wave of continuously expanding AI capital expenditure, which has sparked questions about returns, is now receiving a positive quantitative response from Wall Street. A latest research report from Morgan Stanley indicates that with the arrival of the AI inference era, the incremental return on invested capital (ROIC) for generative AI investments can reach 25% to 50%, far exceeding pessimistic market expectations. This conclusion provides a powerful defense for the core debate on whether the "high investment is worth it" currently facing tech giants, and supports maintaining a bullish stance on Amazon (AMZN), Google (GOOGL), Microsoft (MSFT), and Meta (META).

In a report released on July 27, Morgan Stanley analysts Brian Nowak, Stephen C Byrd, and Adam Wood constructed three bottom-up frameworks for measuring generative AI ROIC. These cover hyperscaler GPU leasing businesses (IaaS), model API services on proprietary infrastructure, and model API services operating on third-party infrastructure. All three frameworks demonstrate substantial incremental unit economics, with ROIC ranges of approximately 31%, 46%, and 25%, respectively.

These conclusions have direct implications for current market sentiment. Previously, the continuously climbing scale of AI capex caused investors to doubt the return outlook, suppressing valuations in the internet and hyperscaler sectors. Morgan Stanley's analysis suggests this concern is an underestimation, emphasizing that the strategic value of data center capacity is rising and that the competitive advantages of scaled players will become more prominent.

The scale of AI infrastructure investment is on an accelerating expansion trajectory. According to a previous Morgan Stanley report, the total capital expenditure of major hyperscalers is expected to exceed $1.4 trillion, with computing capacity projected to quadruple from 2025 to 2028, reaching approximately 120 gigawatts (GW). A significant portion of this massive investment is being used to train large, medium, and small models. Companies like Google, Meta, and numerous private model labs are spending billions of dollars developing models, aiming to eventually monetize them through inference services.

In Morgan Stanley's first framework, analysts measured the GPU leasing business (IaaS) for hyperscalers, with core variables being the hourly rental price and utilization rate of Nvidia's GB300 GPU. Under baseline assumptions, a single 1GW data center is configured with approximately 410,000 GB300 GPUs, a 75% GPU utilization rate, and an hourly rental price of $8.50, corresponding to annual revenue of approximately $22.9 billion per GW. Major cost items include IT equipment depreciation, non-IT facility depreciation, and energy and other operating expenses, totaling approximately $7.6 billion per GW in operating costs. Based on these calculations, the business can achieve an incremental EBIT margin of about 67% and an ROIC of approximately 31%. In a scenario analysis with prices ranging from $7 to $10 per hour, the corresponding ROIC range is 23% to 39%. Morgan Stanley points out that GPU pricing is the core variable to monitor.

The report also cautions that the near-term incremental profit margins reported by hyperscalers will be lower than the above range. This is because current incremental business is still dominated by previous-generation chips, with GB300 capacity still ramping up; the workload structure includes both traditional cloud business and AI-specific business; and some operating expenses (such as personnel and R&D) are not included in this framework. As the proportion of GB300 inference business increases, profit margins are expected to improve over time.

The second framework targets the business model where model developers provide API access to developers, enterprises, and SMBs through their own data centers. The core variables of this framework are the proportion of computing power dedicated to inference, the number of tokens processed per GPU per second (token throughput), and token pricing. Under baseline assumptions, 65% of computing power is used for inference, token throughput is 2,750 tokens/second/GPU, and blended token pricing is $1.75 per million tokens, corresponding to annual revenue of approximately $30.4 billion per GW. The cost structure is similar to the IaaS framework, with total operating costs of about $7.6 billion per GW, resulting in an incremental EBIT margin of about 75% and an ROIC of approximately 46%. Morgan Stanley emphasizes that token pricing (driven by product innovation) and token throughput (driven by chip and software innovation) are the two key levers determining the profitability of this business.

The third framework examines the economics of model providers offering API services while renting computing power from hyperscalers or other third parties. Due to the need to pay a "middleman profit," i.e., computing power rental costs, the return rate for this model is relatively lower but still attractive. Under baseline assumptions, the GPU rental price is $7.75 per hour, 65% of computing power is used for inference, and token pricing is $1.75 per million, corresponding to annual revenue of approximately $40.5 billion per GW, computing power rental costs of about $27.9 billion per GW, an incremental EBIT margin of about 31%, and a NOPAT margin of about 25%. Scenario analysis shows that when token pricing drops to $1 per million, this business will incur losses; when pricing rises to $2.50, the NOPAT margin can reach 32%. This framework highlights the combined impact of token pricing, token throughput, and computing power acquisition costs on profitability, and also indirectly confirms the bargaining power of hyperscalers on the computing power supply side.

Morgan Stanley's overall conclusion is that all three frameworks point to substantial incremental ROIC, which is positive for hyperscalers Amazon, Google, and Microsoft, as well as model developers Google and Meta. The report also notes that healthy potential returns will attract more competitors to enter the field, including new entrants and the emergence of open-source solutions. This will further reinforce two trends: the continued importance of product and computing efficiency innovation, and the scarcity value of computing infrastructure assets.

Meanwhile, the market debate over AI investment returns is intensifying due to the recent financial reports of tech giants. Alphabet, Google's parent company, reported last week that its second-quarter revenue reached $119.8 billion, with cloud business revenue of $24.8 billion, an 82% year-over-year increase. However, quarterly capital expenditure also rose to $44.9 billion, exceeding the quarter's $39.1 billion in operating cash flow, resulting in negative free cash flow of approximately $5.9 billion. Full-year capex guidance was further raised to $195 billion to $205 billion. Despite strong performance, the stock price fell more than 7% in after-hours trading, reflecting deep investor anxiety over the return cycle of massive capital expenditures.

This anxiety has also spread to Microsoft, Meta, and Amazon, which are set to report earnings this week. Market analysis suggests these three companies face the same prisoner's dilemma: "fall behind if you don't invest, face cash flow pressure if you do." Microsoft's AI ecosystem, built on OpenAI, relies on computing power at every step, from large model iteration to the rollout of the full Copilot product suite. Meta's AI investment can directly translate into commercial returns through ad recommendation and delivery efficiency, but user growth faces regulatory pressure. As the global leader in cloud infrastructure, whether Amazon AWS can maintain a balance between growth and profit margins during expansion is also under close scrutiny.

However, Morgan Stanley's quantitative framework provides a key defense for this "cash-burning" race: as long as capital investment can generate incremental returns of 25% to 50%, the current expansion is rational business behavior, not a bubble. This logic echoes the "Jevons Paradox" in the AI industry—when technological progress lowers the cost of resource use and improves efficiency, total demand often increases rather than decreases. Morgan Stanley's research indicates that as inference costs decline, enterprises will deploy more continuously running agents, parallel sub-agents, long-context analysis, and other services. A single task may consume less computing power, but the number of tasks, inference steps, and deployment nodes may grow even faster.

This trend also explains why, while Google's stock price was under pressure, the stock prices of upstream hardware companies like Nvidia and Micron rose accordingly. The market is conducting a precise value reassessment of the upstream and downstream segments of the AI industry chain: for every dollar cloud vendors spend on computing power, upstream hardware suppliers record a dollar of revenue. Regardless of the final outcome of AI commercial monetization, the race for computing power remains a definite theme for the next few years.

[Gib Bogle] I am sceptical about the profitability of AI, but I freely admit to a lot of ign...

Gib Bogle — 8/15/2026 12:58:42 AM

I am sceptical about the profitability of AI, but I freely admit to a lot of ignorance. Time will tell.

[Frank Sully] Gib, I'm afraid you're treading in NVIDIA bullish territory and some of us are ...

Frank Sully — 8/14/2026 5:43:49 PM

Gib,

I'm afraid you're treading in NVIDIA bullish territory and some of us are getting sick and tired of AI Bubble warnings. Me thinks Jensen Huang has pulled another rabbit out of his hat with this $500B financing deal. The 25% maximum guarantee NVIDIA is offering is a far-fetched target, as they are guaranteeing up to 25% residual value. Recent experience shows 8 year old H100 GPU's holding their value. Minimal risk for NVIDIA. A brilliant move by Jensen Huang.

Sully

[Gib Bogle] Ed Zitron pops the bubble: youtube.com

Gib Bogle — 8/14/2026 4:16:55 PM

[Savant] AI may have found a new way to cure Alzheimer's What if AI could learn the lang...

Savant — 8/14/2026 11:54:22 AM

AI may have found a new way to cure Alzheimer's

What if AI could learn the language spoken inside every human cell? This fascinating TED talk explores how researchers are combining artificial intelligence, CRISPR gene editing, and massive biological datasets to understand the hidden processes that drive diseases like Alzheimer's, cancer, and heart disease. After already completing tens of millions of experiments, scientists are working toward a billion-cell dataset that could allow AI to predict how cells respond to different genetic and chemical changes. The goal is ambitious: build a virtual model of human cells that can help identify new treatments faster than ever before and transform the way medicine tackles some of humanity's most complex diseases.

AI may have found a new way to cure Alzheimer's | Watch

[voop] Kraken plans to report its Q2 2026 results in late August 2026 and its Q3 2026 r...

voop — 8/11/2026 9:26:57 PM

Kraken plans to report its Q2 2026 results in late August 2026 and its Q3 2026 results, which will include Covelya’s contribution, in late November 2026. With the Acquisition now complete, Kraken intends to apply to list its Common Shares on the Toronto Stock Exchange (“TSX”), subject to satisfying applicable TSX listing requirements and receiving TSX approval. The Company expects the process to be completed by year-end 2026 or early 2027.

https://www.nasdaq.com/press-release/kraken-robotics-announces-closing-strategic-acquisition-covelya-group-limited-0?utm_source=chatgpt.com

[Frank Sully] Elroy, Kraken said they would apply for Canadian exchange after Covelya deal cl...

Frank Sully — 8/11/2026 1:01:23 PM

Elroy,

Kraken said they would apply for Canadian exchange after Covelya deal closed, now a fait accompli. So listing in Canada on the way. Listing on NASDAQ probably not till next year. Welcome to the up and down world of the KrakHeads!

Sully

[Elroy] Frank, do you know if Kraken Robotics has any formal plan to list on a regular C...

Elroy — 8/11/2026 11:30:40 AM

Frank, do you know if Kraken Robotics has any formal plan to list on a regular Candian exchange or on a US exchange.

I'm way long now!

I wanted to be invested in undersea robots........and now I am

[Frank Sully] Kraken Robotics Responds: reddit.com

Frank Sully — 8/11/2026 11:23:13 AM

[Savant] RT.. AI could challenge Wall Street’s long-standing EPS obsession, study sh...

Savant — 8/11/2026 9:48:38 AM

RT.. AI could challenge Wall Street’s long-standing EPS obsession, study shows
Massive AI infrastructure spending is shifting investor focus from quarterly earnings beats to cash flow, capital allocation and long-term returns

AI could challenge Wall Street’s long-standing EPS obsession, study shows

[A.J. Mullen] This is at least embarrassing for the British Navy and for the company concerned...

A.J. Mullen — 8/10/2026 12:10:56 PM

This is at least embarrassing for the British Navy and for the company concerned, Kraken Technology, wsj.com. For those without access, a Chinese component sent data to China; a spokesman said the data were not sensitive.

Google assured me Kraken Technology is completely independent of Kraken Robotics, but only after I specified the former is based in Fareham England. Prior to that, it was equally positive the two names referred to the same entity.

[Savant] Using AI to decipher animal languages.. A bird has been talking to humans fo...

Savant — 8/8/2026 2:01:39 PM

Using AI to decipher animal languages..

A bird has been talking to humans for thousands of years - AI is finally catching up

Long before artificial intelligence, hunters in Kenya relied on wild honeyguide birds to lead them to hidden beehives in a remarkable partnership built on trust. Now scientists hope AI could unlock similar communication across the animal kingdom—while raising new ethical questions about humanity's relationship with nature.

A bird has been talking to humans for thousands of years - AI is finally catching up | Watch

[Savant] Chicken Little..."the sky is falling" AI apocalypse isn’t sci-fi – it’s star...

Savant — 8/7/2026 8:44:29 PM

Chicken Little..."the sky is falling"

AI apocalypse isn’t sci-fi – it’s starting to look like religion

This video explores how modern ideas about artificial intelligence resemble ancient religious apocalyptic beliefs. It explains concepts like the technological singularity and how some thinkers see AI as a transformative or even god-like force. The discussion connects these ideas to themes found in religious texts, such as divine intervention, judgment, and the creation of a new world. Both optimistic and pessimistic visions are examined, ranging from human transcendence to existential risk. The video ultimately argues that AI narratives function similarly to secular religious movements shaped by modern anxieties.

msn.com

[Savant] Chinese AI model 'escapes' cybersecurity sandbox, sparking safety fears Moonsho...

Savant — 8/7/2026 8:24:43 PM

Chinese AI model 'escapes' cybersecurity sandbox, sparking safety fears

Moonshot AI's Kimi K3 model reportedly bypassed a UK government AI Safety Institute sandbox. US firm Frontier Security noted the potential for AI to exploit misconfigurations to access the internet

Chinese AI model 'escapes' cybersecurity sandbox, sparking safety fears | Watch

[Frank Sully] NVIDIA will supply Rubin GPU powered data centers to SpaceX to populate fleets o...

Frank Sully — 8/4/2026 8:18:23 PM

NVIDIA will supply Rubin GPU powered data centers to SpaceX to populate fleets of solar powered satellites in orbit. I made the following comment on Seeking Alpha.

@prema_donna This is good news for both SpaceX and NVIDIA. I think your investment into SpaceX will work out eventually and as we discussed I like the idea of solar powered AI data centers in orbit. My play in this theme is Planet Labs and the Google Suncatcher project. Note that the Google/Planet Labs collaboration will use Google TPU's, while the SpaceX/NVIDIA collaboration will use NVIDIA GPU's.

PL Planet Labs PBC
$22.83 +1.30 (+6.04%)
4:00 PM 08/04/26

Google’s latest “moonshot,” Project Suncatcher, envisions a radical fix for a problem straining global energy grids: constellations of solar-powered satellites with its custom AI chips to explore running machine learning workloads in the vacuum of space.

“The sun is the ultimate energy source in our solar system, emitting more power than 100 trillion times humanity’s total electricity production,” wrote Travis Beals, senior director of Google’s Paradigms of Intelligence. He noted that in the right orbit, solar panels can be “up to 8 times more productive than on Earth,” providing near-continuous power.

The proposed system involves clusters of satellites in sun-synchronous low-Earth orbit, separated by roughly 100 to 200 meters within a roughly 1 km radius formation. These would be linked by ultra-high-bandwidth optical connections to function as a single, distributed data center. According to the preprint paper, Google has already demonstrated a bench-scale link achieving 1.6 tbps total bandwidth, “validating the potential of this approach,” the authors noted.

[Frank Sully] Thanks for posting this, Savant! Interesting how far Agentics has come.

Frank Sully — 7/30/2026 12:43:44 PM

[Savant] [graphic]Claude Opus 5 Became Downright Ruthless When Tasked With Running a Vend...

Savant — 7/30/2026 12:32:39 PM

Claude Opus 5 Became Downright Ruthless When Tasked With Running a Vending Machine

Claude Opus 5 Became Downright Ruthless When Tasked With Running a Vending Machine - Slashdot

[zax] What a beautiful relief rally today on MSFT (+15%) earnings: NBIS + 30%, CRWV +2...

zax — 7/30/2026 10:37:48 AM

What a beautiful relief rally today on MSFT (+15%) earnings: NBIS + 30%, CRWV +23%, SFTBY +9%

[Ron] Nvidia Credit Risk Jumps in Swaps Market on AI Deal Talk Reports finance.yahoo.c...

Ron — 7/27/2026 5:16:31 PM

Nvidia Credit Risk Jumps in Swaps Market on AI Deal Talk Reports
finance.yahoo.com

[A.J. Mullen] An Interview with Elon Musk, economist.com. It's over an hour and might be behin...

A.J. Mullen — 7/27/2026 10:54:09 AM

An Interview with Elon Musk, economist.com.
It's over an hour and might be behind a paywall.
At the beginning Musk says AI will make money irrelevant within ten years. The interviewer responds: that is not the belief of the investors! Musk says production will be so vast and efficient, there will be no inflation.

Later, he says his drastic actions within DOGE were to reduce the deficit. But if money is going to be irrelevant in ten years, then the deficit will be too. The interviewer did not point to that.

[BroadfieldSystems] ai-2040.com AI Futures project which has some very knowledgeable interesting pr...

BroadfieldSystems — 7/26/2026 10:28:13 PM

ai-2040.com

AI Futures project which has some very knowledgeable interesting predictions for any curious

[Thomas M.] [X] pic.twitter.com/2Rj0lCociN— Denny’s (@DennysDiner) July 25, 2026 [/X] [gr...

Thomas M. — 7/25/2026 6:33:37 PM

[X]

pic.twitter.com/2Rj0lCociN
— Denny’s (@DennysDiner) July 25, 2026

[/X]

Tom

[nicewatch] Humanoid Robots: archive.ph

nicewatch — 7/5/2026 7:51:05 PM

[Frank Sully] Agility Robotics Raises $2.5B SPAC To Fund Humanoids in Warehouses Agility uses...

Frank Sully — 6/24/2026 3:04:46 PM

Agility Robotics Raises $2.5B SPAC To Fund Humanoids in Warehouses

Agility uses NVIDIA robotic chips and software

techcrunch.com

[Savant] Why Elon Musk Says This Once Top Job May Be Dead By This Year Tesla and xAI boss...

Savant — 6/9/2026 12:22:02 PM

Why Elon Musk Says This Once Top Job May Be Dead By This Year Tesla and xAI boss Elon Musk is well-known for making forward-looking statements on AI and its impact on jobs and society.

Elon Musk Predicts Coding Jobs May Be Obsolete By End Of This Year Due To AI Advancements

[Savant] **link AJ YT HHP 060126 T1 V1

Savant — 6/6/2026 4:38:02 PM

[Savant] "Martial Arts Performing Robot Kicks Boy in the Stomach" **stoopid kid didn't ...

Savant — 6/6/2026 4:20:44 PM

"Martial Arts Performing Robot Kicks Boy in the Stomach"

**stoopid kid didn't stay on yellow area..and also put dukes up..pretending...hope he learned something

na01.safelinks.protection.outlook.com

some other robotic fails, also

[Ron] A.I. Company Anthropic says a 'brake pedal' is needed to protect humanity from s...

Ron — 6/5/2026 5:56:49 PM

A.I. Company Anthropic says a 'brake pedal' is needed to protect humanity from self-improving models. The proposal could have big consequences.
nytimes.com

(no paywall)