Status and play🧠🧩 majorana 1⚛️ claude 3.7🔢 protoclone🦿 evo 2🧬 AI civil servants💼 RIP humane⚰️ the handoff🤖🔄👨💼 dark factories🦾 smart pyjamas👕 #2025.08
i/acc: Ineffective Accelerationism 🐌
Welcome to this week's Memia scan across emerging tech and our exponentially accelerating future. So much going on this week, the newsletter pretty much wrote itself… thanks for reading!
(My hybrid human+AI newsletter curation workflow is still *just about* benchmarking1 ahead of the AI competition. Check out Perplexity Deep Research’s output to the query: List and lightly summarise the 50 leading AI and emerging tech news stories from the last week. Won’t be long now until AI is doing better than humans at this task… which I’m looking forward to as it will free me up to do more thinking about the “so what…?” questions.)
ℹ️PSA: Memia sends *very long emails*, best viewed online or in the Substack app.
🗞️Weekly roundup
The most clicked link in last week’s newsletter was the Onion News Network’s disturbingly believable satire: GigSlave.
ICYMI
Another curation of thinking links out on Monday:
✈️☀️Rollcall
Shoutout to the local team at Ōtautahi-based startup Kea Aerospace on the latest test flight of the the Kea Atmos Mk1b solar-powered UAV. Taking off from the Tāwhaki National Aerospace Centre (basically a few sheds in a field!) the wide-winged aircraft flew for 8 hours and 20 minutes, soaring up into the stratosphere and sending back some spectacular images of my backyard, Te Pātaka o Rākaihautū / Banks Peninsula:
Imagine the sky full of these soaring birds carrying planet-observation cameras and internet connectivity nodes… would surely cost a lot less than a constellation of satellites! (And don’t mention space debris risk…)
🧠🧩Status and play
Research from HSE University and the London School of Hygiene and Tropical Medicine has identified 15 core evolutionary motives that drive human behaviour, organised into five categories:
…which they then represented clustered together in a network graph:
“Using network-based psychometric techniques, we were able to observe how motives interrelate. For instance, the motives of Love and Nurture are positioned close to each other in the network, which makes sense from an evolutionary perspective, as caring for offspring enhances their chances of survival. Conversely, the motives of Fear and Curiosity often have opposing effects. Fear keeps us away from danger, but when excessive, it can suppress curiosity, which fosters knowledge and innovation” — Albina Gallyamova, researcher
In particular, the motives Status and Play serve as hub nodes that connect to many other motives in the network visualisation — which the researchers suggest means that they have widespread influence across the wider motivational system.
Obviously — as with all horoscopes pyschometric frameworks — it’s wise to treat these findings with a hefty dose of scepticism until a far larger sample set is analysed… but at the same time it’s a tantalising reductive lens to boil all human behaviour down to 15 weighted parameters…. if humans really are that simple, AI will have no problem with predicting and manipulating us.
(Turning the lens forward - are there different evolutionary motives could we assign to AI, and/or Human+AI?)
⚛️Majorana 1
Microsoft unveiled Majorana 1, a breakthrough quantum chip with million-qubit potential:
CEO Satya Nadella proudly handling the device during his wide-ranging podcast with Dwarkesh Patel.
“Imagine a chip that can fit in the palm of your hand yet is capable of solving problems that even all the computers on Earth today combined could not!”
What sets Majorana 1 apart is its use of topological qubits. These are built using a new material called a “topoconductor” (topological superconductor), which combines a semiconductor (indium arsenide) and a superconductor (aluminum). This hybrid material creates a special state known as topological superconductivity when cooled to near absolute zero and tuned with magnetic fields. In this “new state of matter”, exotic quasiparticles called Majorana fermions emerge. (Named after physicist Ettore Majorana, who theorized them in 1937, these particles are unique because they are their own antiparticles and have properties that make them highly stable against environmental disturbances.)
Microsoft claims is that with this new toplogical qubit technology, practical quantum computers are “years, not decades” away.
Full-on scientific paper published in Nature: Interferometric single-shot parity measurement in InAs–Al hybrid devices.
More accessible (“dumbed down”?) Microsoft explainer video here: friendly scientist faces speaking convincingly to camera…
…but Quantum BS detectors going off in some places:
Firstly, Majorana 1 is a prototype, not a commercial product: a proof of concept showing that topological qubits can be created and measured reliably. Microsoft’s goal to scale this up into a fault-tolerant quantum computer that can run trillions of operations without errors — is still years away.
Some physicists are questioning Microsoft’s quantum claims…whether the data truly indicates the presence of Majorana fermions—or if it could be explained by more conventional Andreev bound states. (Beyond my paygrade but fascinating how fast the peer-review process is moving…)
Fundamentally: if this quantum hardware works, it’s going to accelerate (Microsoft’s hold on) AI. Satya again:
“Quantum is going to be fantastic for anything that is not data-heavy but is exploration-heavy in terms of the state space. It should be data-light but exponential states that you want to explore. Simulation is a great one: chemical physics… biology.
One of the things that we've started doing is really using AI as the emulation engine. But you can then train. So the way I think of it is, if you have AI plus quantum, maybe you'll use quantum to generate synthetic data that then gets used by AI to train better models that know how to model something like chemistry or physics or what have you. These two things will get used together.
…I hope to replace some of the HPC pieces with quantum computers.“
This announcement, combined with Satya’s reticence to overhype “AGI” — and also reports that Microsoft has cancelled leases for "a couple of hundred megawatts" of AI data center capacity in the US … seems to indicate Microsoft is forking its own path on AI from now on, separately from OpenAI.
🤖🔄👨💼The handoff
Tech seer Kevin Kelly delivers a compelling vision of the impending global techno-demographic shift and its economic implications in The Handoff To Bots. He argues that robots will power economic growth while the human population declines, which in turn implies that current economic systems must adapt from human-centric growth to machine-driven progress.
Rather hopefully, he predicts that humans will shift to non-productive roles while synthetic agents (“synths”) will do all the essential economic tasks:
“Our role in the economy is to do all the kinds of things that would not count as productive. Make art, make music, create crazy things because we can, explore the frontiers of reality, and discover new ideas (with the help of genius machines), try stuff, invent new desires we did not know we had, be creative in a different way than machines are. Also, sit with each other when we are sick, have meals with friends –you know, the most important things in the universe to do…
…We are not replacing existing humans with bots, nor are we replacing unborn humans with bots. Rather we are replacing never-to-be-born humans with bots, and the relationship that we have with those synthetic agents and ems, will be highly mutual. We build an economy around their needs, and propelled by their labor, and rewarding their work, but all of this is in service of our own definition of progress and human success.“
(That’s one rose-tinted utopian spin alright …. personally I’m unconvinced, mainly because he assumes that biological humans and technological “bots” will remain distinct entities rather than converging along a hybrid biotech evolutionary path. These transhuman organisms will have their own agency — and the fundamental question for me is whether or not they end up competing for resources in the same evolutionary niches as humans.)
📈The week in AI
Another huge week of AI news and announcements. Firstly, a look at AI’s wider impacts:
🪖Israel uses US-made commercial AI for war
An in-depth Associated Press investigation reveals that the Israeli military's use of Microsoft and OpenAI technology increased 200-fold after the October 7, 2023 Hamas attack. In particular, the report finds that:
The Israeli military employs commercial AI to rapidly identify targets by analysing intelligence, communications, and surveillance data
Data storage on Microsoft servers doubled to 13.6 petabytes between October 2023 and July 2024
While officials claim AI makes targeting more accurate and effective, the investigation uncovered instances where AI systems contributed to civilian deaths through translation errors, misidentifications, and confirmation bias
Microsoft maintains deep ties with the Israeli military through a US$133 million contract, making Israel its second-largest military customer globally after the US.
In particular, AP reports:
“The Israeli military uses Microsoft Azure to compile information gathered through mass surveillance, which it transcribes and translates, including phone calls, texts and audio messages, according to an Israeli intelligence officer who works with the systems…
...About a year ago OpenAI changed its terms of use from barring military use to allowing for “national security use cases that align with our mission.”
…Google and Amazon also provide cloud computing and AI services to the Israeli military under “Project Nimbus”, a $1.2 billion contract signed in 2021 when Israel first tested out its in-house AI-powered targeting systems…”
Palantir, Cisco and Dell are also mentioned in the article.
The shocking drone footage in the article of the completely destroyed townships in Gaza speaks for itself of the inhumanity which has been enabled to happen since 2023. The casualty numbers also speak loudly:
But will governments and companies withdraw their business from these companies on moral grounds? (And if so, who would they buy from instead? Show me a US company which would dare to differentiate themselves on ethical grounds in the current political environment!?)
(And let’s be honest, would open-source AI change any of this behaviour or result in less genocidal outcomes? Probably not…)
🏭Data centre pollution costs US public health US$5.4 billion
Research from UC Riverside and Caltech asserts that pollution from Big Tech's datacentre expansion is causing significant public health costs, estimated at US$5.4 billion over the past five years, disproportionately affecting lower-income communities. Of course, it goes without saying, the projected AI infrastructure expansion will significantly increase energy consumption and related pollution.
Google, Meta and Microsoft pushed back on the report’s methodology, pointing to “clean energy” purchases to offset emissions. But:
“Unlike carbon emissions, the health impacts caused by a data centre in one region cannot be offset by cleaner air elsewhere” — Shaolei Ren, associate professor at UC Riverside.
Case in point, xAI’s Colossus data centre in Memphis, currently applying to make its temporary gas turbines permanent, against vocal community concerns on air pollution.
📊AI industry news
Skimming across the wires:
⚰️RIP Humane
The Humane Ai Pin will cease functioning this week on February 28, as HP acquires Humane's assets (mostly patents) in a $116 million deal. It was weird while it lasted.
Humane no more 💭Thinking Machines
Thinking Machines Lab is a new AI research and product startup founded by former OpenAI CTO Mira Murati and other veterans from OpenAI, Character.ai and other leading labs.
“While AI capabilities have advanced dramatically, key gaps remain. The scientific community's understanding of frontier AI systems lags behind rapidly advancing capabilities. Knowledge of how these systems are trained is concentrated within the top research labs, limiting both the public discourse on AI and people's abilities to use AI effectively. And, despite their potential, these systems remain difficult for people to customize to their specific needs and values. To bridge the gaps, we're building Thinking Machines Lab to make AI systems more widely understood, customizable and generally capable.“
Safe as houses (?)
Safe Superintelligence, the AI startup co-founded by former OpenAI co-founder Ilya Sutskever, secured over US$1 billion in funding at a staggering US$30 billion valuation — despite having no product, revenue, or immediate plans to sell anything. Even the FT speaks in memes now:
FT
Alibaba commits US$38 billion to cloud and AI expansion Aping its US competition, China’s Alibaba Group announced a landmark 380 billion yuan (US$38 billion) investment in cloud and AI infrastructure over the next three years — the largest commitment by a Chinese private enterprise in this sector and exceeding the company's total investments in these areas over the past decade.
💼 Shenzhen deploys AI civil servants In the public sector, Shenzhen has deployed "AI civil servants" developed on DeepSeek's platform, becoming the first city in Guangdong province to implement this technology citywide. According to the writeup, 70 AI assistants now handle 240 government service scenarios across multiple departments reducing document review time by 90% with 95% accuracy.
(Which Western bureaucracies are experimenting with this??)
Sofa talk Google Deepmind’s Demis Hassibis and Anthropic’s Dario Amodei were forced to share a very small sofa (surely intentional!) while sharing their concerns for AGI Safety on stage with The Economist’s Zanny Minton-Beddoes (who had a bigger sofa all to herself!)
Amazing how even the most powerful men in AI are still socially awkward meat puppets in physical reality.
🆕 AI releases
🔢Claude 3.7
Anthropic finally unveiled Claude 3.7 Sonnet, their latest AI model, featuring hybrid reasoning and improved coding abilities. (What is it with Anthropic’s version numbering? The number 4 is unlucky in Chinese culture… could it be…)
Anyway, Claude 3.7 is another incremental advance on Claude 3.5 Sonnet (New) which was arguably still the SOTA up to this week. Naturally, benchmarks go up:
And most impressively, pricing remains unchanged despite the upgraded capabilities.
Claude 3.7 Sonnet also follows OpenAI, DeepSeek and Grok by introducing “reasoning” capabilities in the form of “Extended Thinking Mode”:
“Extended thinking” is yet another implementation of test-time-compute which allows Claude more time and cognitive effort to tackle complex problems. The Anthropic team took the design decision to clearly show its visible thought processes — in their writeup they say this primarily enhances trust and “alignment”… but is also just downright fascinating to watch. (And not without novel security attack vectors either…)
This graph shows how 3.7 Sonnet outperforms 3.5 Sonnet as more “reasoning” steps are applied.
Anthropic They even benchmarked it on playing Pokémon!
Anthropic Also, introducing Claude Code:
“Claude 3.7 Sonnet shows particularly strong improvements in coding and front-end web development. Along with the model, we’re also introducing a command line tool for agentic coding, Claude Code. Claude Code is available as a limited research preview, and enables developers to delegate substantial engineering tasks to Claude directly from their terminal.“
Meanwhile last week’s new kid on the block, Grok 3 from xAI, has been settling into the race, with a full rollout to X users on the free tier.
“Here are the major practical-level takeaways so far, mostly from the base model since I didn’t have that many tasks calling for reasoning recently, note the sample size is small and I haven’t been coding:
Hallucination rates have been higher than I’m used to. I trust it less.
Speed is very good. Speed kills.
It will do what you tell it to do, but also will be too quick to agree with you.
Walls upon walls of text. Grok loves to flood the zone, even in baseline mode. A lot of that wall is slop but it is very well-organized slop, so it’s easy to navigate it and pick out the parts you actually care about.
It is ‘overly trusting’ and jumps to conclusions.
When things get conceptual it seems to make mistakes, and I wasn’t impressed with its creativity so far.
For such a big model, it doesn’t have that much ‘big model smell.’
Being able to seamlessly search Twitter and being in actual real time can be highly useful, especially for me when I’m discussing particular Tweets and it can pull the surrounding conversation.
It is built by Elon Musk, yet leftist. Thus it can be a kind of Credible Authority Figure in some contexts, especially questions involving Musk and related topics. That was quite admirable a thing to allow to happen. Except of course they’re now attempting to ruin that, although for practical use it’s fine for now.
The base model seems worse than Sonnet, but there are times when its access makes it a better pick over Sonnet, so you’d use it. The same for the reasoning model, you’d use o1-pro or o3-mini-high except if you need Grok’s access.“
More pressingly Grok 3 is scaring the pigeons with its apparent lack of guardrails. The xAI team are clearly unencumbered by concepts such as red-teaming:
And… several hasty updates of Grok 3’s system prompt (easily extracted) seemingly reveal a thin-skinned champion of “free speech”:
(According to senior XAI engineer Igor Babuschkin, the System Prompt change was made by a new XAI employee who joined from OpenAI… the matter was resolved… and the System Prompt changed… apparently to this, after Grok 3 started saying Donald Trump deserved to die…)
(In all probability what seems to be happening here is sleep-deprived AI developers making live changes to production to test things out on the world’s largest GPU cluster…. yikes…)
Uncensored DeepSeek-R1
Perplexity AI open-sourced R1 1776, a DeepSeek-R1 reasoning model post-trained to remove Chinese Communist Party censorship while maintaining high reasoning capabilities. Gotta love open-source.
DeepSeek open-sourcing 5 code libraries
DeepSeek announced plans to release five open-source code repos over five consecutive days starting this week, positioning itself as a small team dedicated to AGI exploration. So far:
🥼 AI research
Magma from Microsoft research, a foundation model serving multimodal AI agentic tasks in both the digital and physical worlds.
Hugging Face continue to maintain a completely open-source Ultra-Scale Playbook for anyone wanting to train their own LLM:
🧬Evo 2 is a massive AI foundation model for biomolecular sciences, built by a research team from Arc Institute, NVidia, Stanford University and other institutions, providing insights into DNA, RNA and proteins across diverse species. Evo 2 is trained on an enormous dataset of nearly 9 trillion nucleotides — the building blocks of DNA and RNA — and “represents a major milestone for generative genomics”:
“…Beyond its predictive capabilities, Evo 2 generates mitochondrial, prokaryotic, and eukaryotic sequences at genome scale with greater naturalness and coherence than previous methods. Guiding Evo 2 via inference-time search enables controllable generation of epigenomic structure, for which we demonstrate the first inference-time scaling results in biology. We make Evo 2 fully open, including model parameters, training code, inference code, and the OpenGenome2 dataset, to accelerate the exploration and design of biological complexity“
Think generative AI but for life. (Now imagine this being included in your $20/month GPT-7 subscription together with access to a molecular printer in 2027...)
NVidia AI solves decade-long superbug mystery in just two days
The BBC reports how Google's AI tool "co-scientist" solved a complex superbug antibiotic resistance research problem in just 48 hours — the same problem that had taken scientists at Imperial College London a decade to investigate. Imperial College London Professor José Penadés of the research team is optimistic about the implications:
"I feel this will change science, definitely…I'm in front of something that is spectacular, and I'm very happy to be part of that.,,It's like you have the opportunity to be playing a big match - I feel like I'm finally playing a Champions League match with this thing."
🧑🤝🧑➕🤖⬇️When Humans + AI < AI
A comprehensive meta-analysis of 106 experimental studies examining human-AI collaboration reveals several key findings:
On average, human-AI combinations performed worse than either humans or AI working alone, though they did outperform humans working independently.
Human-AI combinations perform better in creative tasks than decision-making tasks
AI outperforming humans alone leads to performance losses when combined
Vaccaro, M., Almaatouq, A. & Malone, T. When combinations of humans and AI are useful: A systematic review and meta-analysis. Nat Hum Behav 8, 2293–2303 (2024). https://doi.org/10.1038/s41562-024-02024-1
AI creates bizarre but superior wireless chips in hours
Princeton Engineering and Indian Institute of Technology researchers have demonstrated AI's ability to revolutionise wireless chip design, achieving in hours what typically takes humans weeks to complete…but sometimes the AI-designed chips are so weird that 'humans cannot really understand them'.
Image: Princeton University AI system aims to save coral reefs from extinction Link
Australian researchers are developing a real-time global monitoring system to protect coral reefs using AI, remote sensing, and data integration technologies:
🔮[Weak] signals
Parsing the signals from the noise across tech that isn’t AI…
📱💻Consumer tech
M4 MacBook Air launching in March, foldable iPhone coming 2026?
Apple is set to launch its new M4-powered MacBook Air in March 2025, with both 13-inch and 15-inch models joining other M4 chip devices in the lineup. Also rumours that a foldable iPhone is planned for 2026, which will shake up the form factor a bit:
Concept Render - Apple Insider Huawei launches world's first tri-fold smartphone globally Link
Meanwhile Huawei has launched the Mate XT globally. This is the world's first tri-fold smartphone that offers three distinct display configurations:
1) a fully unfolded 10.2-inch tablet mode at 2,232 × 3,184 pixels
2) a partially folded 7.9-inch dual-screen at 2,232 × 2,048 pixels
and 3) a completely folded 6.4-inch smartphone mode at 2,232 × 1,008 pixels...
Premium pricing at €3,499 establishes a new ceiling for smartphone market segments… Definitely not my kind of luxury accessory (imagine having 3 cracked screens instead of only one?!)
Huawei Mate XT. That gaze...
🧮RISC-Y
AheadComputing, a new chip startup founded by former Intel executives, has secured US$21.5 million in seed funding to develop 64-bit RISC-V-based CPU processors targeting AI computing challenges. (Reminder: the open-source RISC-V reduces dependency on Intel/AMD x86 and ARM's proprietary designs.)
Future of news?
Noosphere is a new premium news platform launched by a group of experienced foreign correspondents, aiming to revitalise independent journalism through a TikTok-inspired mobile-first experience. US$20 monthly... Founder and CEO Jane Ferguson, a former war correspondent for PBS, explains the origins to Semafor:
“I had essentially built an incredibly successful career in an industry that was collapsing, which is a very weird place to be, where you just collect trophies while the audiences are getting smaller and the finances are not working…And it became apparent to me that this was actually terminal…What we really wanted to do is build the architecture for the best journalists to enter the content creator economy.”
(The existential question for journalism: will humans still be involved in collecting and generating original news any more … or will Grok 3 and its competitors just scour the wires for signals from now on, parsing these into actionable intelligence in real time…)
🦿More humanoids
Following on from last week’s humanoid roundup, even more grist to the mill:
Figure Helix US robotics startup Figure launched Helix: A Vision-Language-Action (VLA) Model for Generalist Humanoid Control.
Helix The model’s architecture owes a bit to Daniel Kahnemann, incorporating Vision Language Model(VLM) and a “visuomotor” policy:
Helix is a first-of-its-kind "System 1, System 2" VLA model for high-rate, dexterous control of the entire humanoid upper body.
Prior approaches face a fundamental tradeoff: VLM backbones are general, but not fast, and robot visuomotor policies are fast but not general. Helix resolves this tradeoff through two complementary systems, trained end-to-end to communicate:
System 2 (S2): An onboard internet-pretrained VLM operating at 7-9 Hz for scene understanding and language comprehension, enabling broad generalization across objects and contexts.
System 1 (S1): A fast reactive visuomotor policy that translates the latent semantic representations produced by S2 into precise continuous robot actions at 200 Hz.
This decoupled architecture allows each system to operate at its optimal timescale. S2 can "think slow" about high-level goals, while S1 can "think fast" to execute and adjust actions in real-time.
Here’s the POC: two robots collaborating to pack the groceries. Slow but watch this accelerate from here…
Neo Gamma
Staying with the household theme, 1X has unveiled Neo Gamma, a humanoid home robot designed to perform household work while providing companionship.
World's first humanoid robot front flip
Chinese robotics company Zhongqing Robotics (EngineAI) claims to be the world’s first to achieve a front-flip. (Priced at US$13,700, this open-source PM01 robot demonstrates how quickly advanced humanoid technology is becoming commercially accessible.
🦿Protoclone
A preview from Polish / UK startup Clone Robotics unveiled Protoclone, a twitching humanoid robot with artificial muscles. Clone's Myofiber technology mimics human musculoskeletal systems for more natural movement. The company plans to go into production of 279 Clone Alpha robots begins this year
(Video: Clone Robotics). Here’s a video of the forearm and hand in action:
🪙Crypto
Mixed signals this week as the crypto space slowly matures…
Coinbase pushes for clear crypto legislation
Coinbase, the largest US cryptocurrency exchange, released a legislative blueprint aiming to influence Washington's approach to crypto regulation under the Trump administration. The blueprint proposes legislation to:
clearly define crypto tokens and transactions
establish a Stablecoin framework
shifting US crypto asset oversight to the Commodity Trading Futures Commission (CFTC) — which would completely overhaul regulation of the US$3.3 trillion market
legal protections for blockchain developers and self-custody wallet rights
Crypto money backed Trump to get in… but will the US financial establishment buy in?Bybit crypto exchange hit by massive US$1.5 billion hack Link
Cryptocurrency exchange Bybit fell victim to a sophisticated hack resulting in the theft of approximately US$1.5 billion worth of Ethereum, making it one of the largest digital heists in history...it’s still the Wild West out there… would legislation change this?
Cardano adopts constitution and achieves full blockchain decentralisation
Flying under the radar somewhat, Cardano (ADA) achieved a significant milestone with the ratification of its community-ratified blockchain “constitution”, implemented this week on February 23.
Following last December’s Constitutional Convention jointly hosted in Buenos Aires and Nairobi, the Cardano community ratified the new constitution, which eliminates central control points.
Drafting Cardano’s Global governance framework involved 1,800 contributors from 50 countries
This week’s implementation (effectively uploading the constitution as code) sets a new precedent for blockchain governance, with ADA holders directly shaping network evolution. Cardano has entered the “Voltaire Era”:
(Congratulations to friend and Memia stalwart @Yojoflo who attended the convention and has been an active participant in the efforts getting this across the line). Cardano now has a treasury of 1.7 billion ADA (~US$1.36Bn) which can be allocated to projects according to decentralised, community-driven processes.
☢️⚡Nuclear power
Two stories catching my eye this week:
The fusion race is on
France's CEA WEST Tokamak reactor has set a new fusion record by maintaining a stable plasma reaction for 1,337 seconds (over 22 minutes) on February 12, surpassing China's January achievement of 1,066 seconds by 25%. Temperatures reached 50 million degrees celsius inside the reactor!
The WEST Tokamak reactor Floating nuclear plants?
Meanwhile startup Core Power unveiled plans to mass produce floating nuclear power plants (FNPPs) using Generation 4 molten salt reactors and modular shipbuilding techniques, targeting deployment off the US coastline expected within a decade.
Core Power 
Core Power
(What could possibly go wrong…?)
🛸Drones
An underwater theme…
Underwater drone swarm to monitor 3D-printed reefs
A groundbreaking project is deploying autonomous underwater vehicles (AUVs) to monitor artificial reefs off Cyprus's coast, aiming to revitalize marine ecosystems in barren sea areas... at scale:
"Our goal is to make a system under which six of our mini-drones can do the same work for the price of one conventional drone,“
One of the submersible drones Bionic fish robot swims in multiple modes like real mackerel Link
Researchers at the Shenyang Institute of Automation have developed a soft robotic fish capable of performing four distinct swimming modes:
Biomimetic robots (eg the Protoclone above) are a trend I’m watching closely… imagine swarms of insect-size robo-bees that look just like real bees… changes the world pretty radically.
🦾Dark factories
The “dark factory”, where manufacturing is wholly automated, no human workers are involved, so the lights are off. This is now reality in China:
(Video via @XH_Lee23)
🪟Smart windows
Researchers from Nanjing University of Aeronautics and Astronautics have developed a flexible dual-band electrochromic window that can reduce building energy consumption by up to 20% by optimising thermal regulation in many different climates. 20% is a BIG number!
☁️Cloud brain
The chairman of Chinese company Guide Infrared claimed a significant lead over Neuralink in brain-computer interface (BCI) technology at a recent conference, with 20X more channels than Neuralink and with bi-directional capabilities. Guide Infrared’s 100 million yuan (US$14 million) “Cloud Brain Project” aims to support global medical scientists in using Guide Infrared’s BCI technology for clinical applications and research.
Not a lot more information out there right now. This article from last year sheds more light on the tech: Chinese researchers develop world-leading brain-computer interface chips.
🌙Race to Luna
A roundup in Techspot of the new private space race to establish a sustainable lunar economy. How long until the artist’s impression above is realised - 2030? 2035? (And wouldn’t everything in the greenhouse die from radiation poisoning…?!)
👕Smart pyjamas
Cambridge University researchers have developed a smart pyjama top that diagnoses sleep disorders. (That’s a comfier form factor than a Whoop wrist band or Oura ring!)
⏳ Zeitgeist
Trying to make sense of the world outside tech…
🌊💩Flooded Zone
Perhaps its just general numbness to the constant attention-hijacking … but nothing Trump, Musk and their acolytes can do is shocking any more after their first month in office. The globe is still lurching back and forth trying to absorb the whiplash of the Trump administration’s violently erratic policy towards Europe and the international order:
US and Russian foreign ministers met in Riyadh aiming to divvy up the spoils in Ukraine… but Ukraine and Europe not in the room. Russian and US diplomatic relations are being restored.
Analysis I’ve heard is that Trump’s wooing Russia is aimed to counter China’s influence in Moscow… but that would imply a rationality that isn’t immediately apparent on the surface.
Ukraine agreed to a watered-down minerals deal with the US after rejecting Trump’s initial demands for US$500bn in potential revenue from exploiting the resources.
Make Europe Great Again (MEGA): immediately following a decisive2 election win, Germany’s new Chancellor-elect Friedrich Merz said:
“My absolute priority will be to strengthen Europe as quickly as possible so that we can achieve, step by step, real independence from the US"
Actions, not words, will tell.
Meanwhile in my part of the world, three Chinese warships sailed down the Tasman Sea between Australia and New Zealand, carrying out a “live firing” exercise which caused civilian planes to change their flight paths. The ships are currently 218 nautical miles off Tasmania.
(I covered analysis and reaction to this in Monday’s Mind Expanding post… not a lot AoNZ can do if China did choose to use force).
China's Jiangkai-class frigate Hengyang. Photo: AFP / Australian Defence Force (Who wins from all this? Well, arms manufacturers, for one…. If AI was running the planet would it allocate 3% of global GDP on weapons for humans to fight their little wars amongst themselves…? Or just take out the most aggressive individuals and use the money to build more data centers? I think I know the answer.)
🍎🇺🇸Apple reshoring (for now)
Apple announced a US$500 billion investment in the US over the next four years, along with plans to hire 20,000 new employees, in what appears to be a strategic move to avoid potential Trump administration tariffs on Chinese imports. The four year timeframe is telling.
💱🇸🇦Saudi Arabia unveils new currency symbol
Ascendant Saudi Arabia unveiled a new currency symbol for the riyal, joining the exclusive group of nations / blocs with dedicated currency symbols like the dollar ($), pound (£), yen (¥) and Euro (€). The symbol draws from Arabic calligraphy, incorporating the letters ر (Ra), يَا (Ya), and ل (Lam) in a geometric form.
Currently not implemented in Unicode font support as far as I can tell, here’s what it comes out like (Unicode Hex Value: FDFC): ﷼
In brief:
A roundup of other interesting non-tech stories I’ve caught this week:
Sudanese army recaptures Obeid In the bloodiest conflict on the planet right now, the Sudanese Army recaptured El-Obeid, a strategically significant southern city, from the paramilitary Rapid Support Forces (RSF). This could increase humanitarian aid to millions displaced in the war, but the RSF's parallel government threatens further instability in already devastated Sudan. Again, who profits from this? (A: Arms manufacturers!)
X-37B sends a selfie from High Earth Orbit
The X-37B on the ground The US Space Force shared a photo from its secretive autonomous X-37B space plane during its seventh mission, showing the vehicle above Earth while performing an aerobraking manoeuvre... no-one’s quite sure what it’s doing up there — its previous mission lasted 908 days.
BlackRock summit aims to tackle America's retirement savings crisis
BlackRock and the Bipartisan Policy Center are organizing a March 12 summit in Washington to address what they call an "impending retirement crisis" in America.
Fear of retirement insecurity exceeds fear of death for over half of Americans surveyed, with only 33% of Americans having retirement savings.
🚅Canada plans 185-mph rail network connecting major cities
Alto Canada is in the early planning stages for "Alto," a high-speed rail network that will be the country's largest infrastructure project ever, connecting 18 million Canadians between Toronto and Québec City. BUT… the six-year co-development phase requires C$3.9 billion before construction can begin and Canada is heading to the polls imminently…
Baltimore's new bridge is designed to last a century
More infrastructure: Baltimore's new Francis Scott Key Bridge replacement, designed by engineering firm Kiewit and due to be completed by 2028, will feature a cable-stayed design with significant safety improvements following the spectacular 2024 collapse when a container ship hit it. The new bridge has a target 100-year lifespan.
Corporations build nuclear-proof bunkers to protect data and executives
Large corporations, particularly cryptocurrency firms, are increasingly investing $millions in underground bunkers to protect both their data centres and executives amid rising geopolitical tensions and threat of nuclear conflict. And AGI:
Asteroid's impact risk shifts from Earth to Moon
Finally, Asteroid 2024 YR4, recently causing concern with a record-high 3.1% impact probability for 2032, has now been downgraded to a minimal threat with a higher chance that it will hit the Moon than the Earth. Phew
The images in which asteroid 2024 YR4 was first identified by the Asteroid Terrestrial-impact Last Alert System (ATLAS)
🎭Meme stream
A productive week of prospecting in the meme-mines…!
😂Ignore all previous instructions
Elon Musk’s (ketamine-addled?) decree-by-tweet that all US federal workers must send a weekly email with 5 bullet points detailing what they’ve achieved this week has seen some pushback. This is the best response so far:
🤔Choose your thinkboi
I live a long way from Silicon Valley but I get every one of these references…
🎬How to make a Porsche ad
Behind the scenes with filmmaker László Gaál. Watch right to the end… clever. (🎩top spotting Geoff D)
👓A history of AR headsets
Every VR/AR headset in history, morphed into a timeline. Neatly done by tapwithus.com. (via Dovid Schick).
🗳️Vote Not Shit
It’s election year in Australia… vote for your local Not Shit candidate!
🐌i/acc
The latest Silicon Valley movement: Ineffective accelerationism …(shhh…that’s the joke!)
🐬Pod cast
Finally, sharing this short video of a playful pod of dolphins accompanying me last weekend on the Bluebridge Te Moana-o-Raukawa / Cook Strait ferry crossing just off Miramar. Made my day.
🙏🙏🙏 Thanks as always to everyone who takes the time to get in touch with links and feedback. What a week!
Namaste
Ben
As measured by the highly scientific BenBench™ suite of (finger-in-the-air) tests…
As decisive as MMP can ever deliver, anyway…. reasons for the 5% threshold clear in this case.