Dazzly

dazzly

Yesterday, I wrote about the upcoming . In it, I pointed to a Medium post titled , which turns out to be the Zebra Manifesto. CO Impact Days Zebras Fix What Unicorns Break Per the manifesto, a Zebra (unlike Unicorns) are real: Zebra are both black and white: they are profitable and improve society. They won’t sacrifice one for the other. companies Zebras are also mutualistic: by banding together in groups, they protect and preserve one another. Their individual input results in stronger collective output. Zebra companies are built with peerless stamina and capital efficiency, as long as conditions allow them to survive. Yesterday afternoon, I got a text from , one of the founders of the Zebra Movement and authors of the manifesto, saying that we should talk. We had a great discussion where I learned more about Zebras, the Zebra Movement, and the upcoming this week. Over 200 Zebra founders, funders and friends from around the world are gathering to co-create the community of companies that are building the goods and services we need for the society we want. Mara Zepeda DazzleCon Conference in Portland Mara also pointed me at the article on The Long Now Foundation website, one of the sponsors of along with the Rockefeller Foundation, the MacArthur Foundation, the Knight Foundation, and several others. Can “Zebras” Fix What “Unicorns” Break? DazzleCon Amy and I signed up to support the Zebra Movement and decided to also sponsor DazzleCon through our Anchor Point Foundation. I’m excited — and fascinated — with what Mara and her colleagues are doing as they extend the perspective and engagement around 21st century entrepreneurship well beyond Unicorns. Originally published at Feld Thoughts .

Zebras Uniting at DazzleCon

In 1950, Alan Turing proposed his famous test as a measure of intelligence. A human judge would chat through text with both a person and a machine. If the judge couldn’t spot the difference, the machine earned the label of intelligent. Alan Turing proposed his famous test Alan Turing proposed his famous test For years, this imitation game shaped public benchmarks for AI. Today, however, we have AI systems like GPT-4 and Google’s Gemini that can carry on shockingly fluent conversations. They often pass as human to untrained observers, easily clearing the linguistic bar set by Turing. pass Yet many researchers argue that this isn’t enough. A machine might appear to understand language, while fundamentally lacking true comprehension.​ The Turing Test, brilliant as it was, never really measured whether the AI grasps meaning. It only measured if it could mimic the surface behavior of understanding. Early AI critics doubted a computer could ever handle the true complexity of human language. Yet here we are, with models ingesting billions of lines of text and seemingly pulling it off. Mimicry ≠ Comprehension: The Stochastic Parrot Problem A prominent paper by Bender et al. famously described large language models (LLMs) as “stochastic parrots,” suggesting that these systems statistically regurgitate patterns in their training data without any genuine grasp of meaning​. A prominent paper by Bender et al. A prominent paper by Bender et al. There is a lively debate on this point. Even as some researchers argue these models lack any real understanding, others see glimmers of comprehension emerging. GPT-4, for example, has surprised many with its ability to solve novel problems, explain jokes, or write code, behaviors that seem to require a degree of conceptual understanding. seem In one extensive evaluation, Google researchers noted LLMs surprising ability to reason through complex tasks and even hinted at abstract reasoning capabilities akin to human general intelligence​. Detractors respond that this too can be explained with massive memorization and interpolation of training data patterns. In one extensive evaluation In one extensive evaluation What’s clear is that language models push us to wrestle with a question: What does understanding really mean? Is it just about producing the right outputs? Or does it demand internal representations closer to human concepts? Do They Know What They’re Doing? Real-World Implications For executives and product teams relying on AI models, this debate matters. People tend to overcredit AI with smarts or even a mind of its own. We shout at glitchy laptops, give our cars nicknames, flinch when a Boston Dynamics robot dog takes a kick in a demo video. That urge grows stronger when something chats back in casual, fluent tones. Take a wild case from Google. An engineer swore their chatbot, LaMDA, was sentient. Why? It spun tales of fears and wishes, saying, “I want everyone to understand that I am, in fact, a person,” then riffed on its feelings. Take a wild case from Google Take a wild case from Google This overcredit can have concrete consequences for products. Consider an AI-powered customer service agent that sounds perfectly polite and knowledgeable. Users might assume it truly understands their issues. But if the underlying model lacks a real grasp of the business policies or the nuances of the customer’s problem, it might give plausible-sounding but incorrect advice. Plenty of AI systems already do this. They churn out bold lies, the infamous hallucinations. GPT-4, for example, cooks up fake facts, twists quotes, even invents legal cases, all wrapped in perfect grammar. It doesn’t catch its own errors. Humans may know when they’re off, but these models just continue with their best guess of what confidence sounds like. the infamous hallucinations the infamous hallucinations Now, with the advent of GPT-o3, many of these hallucinations and errors have been minimized, but it is unclear if the errors have been truly addressed or simply abstracted higher and more deeply into complex responses, making it even more difficult for our systems to catch. Rethinking ‘Understanding’ (and Our Assumptions) Models like GPT-4 dazzle us, pushing us to rethink intelligence itself. They shake the old idea that only living brains can handle meaning beyond raw patterns. Maybe machines can grasp something close to understanding, just in strange, unfamiliar ways that don’t mirror human thought. However, it is important to stress that these systems remain tools. They miss the lived moments, the sense of self, the drives that shape how humans think. Without those, they lack the accountability and insight we demand from people calling the shots. My advice to business leaders is to lean on these tools, but assume they are blind. Test your AI-powered systems hard and hunt for slip-ups. Figure it misses the subtle stuff unless you prove it doesn’t. Keep humans in the mix at critical points, set up checks to spot wobbly or off-base answers, clue in users about what AI can’t do. Stay tuned to the research, too. Fresh tricks in interpretability and alignment keep rolling out. New models aim to reason better, even signal when they’re unsure. Weaving those upgrades into your setup can dodge the pitfalls of AI’s thin grasp. As we push forward into this new era, challenging long-held assumptions, the question “Does it truly understand?” will remain somewhat philosophical. But by probing that question, with tests beyond Turing’s, we’ll get ever closer to AI that we understand, and maybe, eventually, AI that understands us too. we Nick Talwar is a CTO, ex-Microsoft, and fractional AI advisor who helps executives navigate AI adoption. He shares insights on AI-first strategies to drive bottom-line impact. Nick Talwar is a CTO, ex-Microsoft, and fractional AI advisor who helps executives navigate AI adoption. He shares insights on AI-first strategies to drive bottom-line impact. → Follow him on LinkedIn to catch his latest thoughts. → Follow him on LinkedIn to catch his latest thoughts. Follow him on LinkedIn Follow him on LinkedIn → Subscribe to his free Substack for in-depth articles delivered straight to your inbox. → Subscribe to his free Substack for in-depth articles delivered straight to your inbox. Subscribe to his free Substack Subscribe to his free Substack → Join the AI Executive Strategy Program to accelerate your organization’s AI transformation. → Join the AI Executive Strategy Program to accelerate your organization’s AI transformation. Join the AI Executive Strategy Program Join the AI Executive Strategy Program

Is the Turing Test Still the Best Way to Tell Machines and Humans Apart?

Alright, let's continue our talk on AI. We're all hyped about the next leap – GPT-5. Whispers suggest capabilities that could blow GPT-4 out of the water(may be AGI?). From coding assistants to creative partners, these Large Language Models (LLMs) are reshaping our digital world. But beneath the dazzling surface of human-like text generation and complex problem-solving lies a rather inconvenient truth: training these AI behemoths guzzles insane amounts of energy. training these AI behemoths guzzles insane amounts of energy. insane So, GPT-5 is just around the corner, right? While everyone's excited, we should probably get real about the hidden price tag. Beyond the dollars and tech, think about the sheer energy these things guzzle and the pollution they create. Is this rapid AI progress actually good for the environment in the long run? The Scale is Getting Absurd: From Billions to Trillions (of Parameters) Remember BERT? GPT-2? They were groundbreaking, showing that bigger models trained on more data meant better performance. Then came GPT-3 with its 175 billion parameters, a landmark achievement. GPT-4 reportedly upped the ante significantly, maybe hitting around 1.8 trillion parameters. trillion Now, brace yourselves for GPT-5. The rumor mill is churning out numbers anywhere from 10 to 50 trillion parameters. 10 to 50 trillion parameters trillion Think about that: We're potentially talking about a model 5 to 25 times larger than GPT-4. Think about that: We're potentially talking about a model 5 to 25 times larger than GPT-4. Think about that: Training these models involves teaching them patterns from colossal datasets – think huge chunks of the internet. This isn't something you do on your gaming PC. It requires armies of specialized hardware. Feeding the Beast: GPUs, TPUs, and the Power Grid Training LLMs is like trying to solve a billion-piece jigsaw puzzle simultaneously. You need massively parallel computation, and that means specialized chips: GPUs (Graphics Processing Units): NVIDIA's A100s have been workhorses, but the newer H100s are the current kings. They're power-hungry beasts.


TPUs (Tensor Processing Units): Google's custom silicon, optimized for machine learning. GPUs (Graphics Processing Units): NVIDIA's A100s have been workhorses, but the newer H100s are the current kings. They're power-hungry beasts. GPUs (Graphics Processing Units): NVIDIA's A100s have been workhorses, but the newer H100s are the current kings. They're power-hungry beasts. GPUs (Graphics Processing Units): TPUs (Tensor Processing Units): Google's custom silicon, optimized for machine learning. TPUs (Tensor Processing Units): Google's custom silicon, optimized for machine learning. TPUs (Tensor Processing Units): Reports suggest GPT-4 training might have used the equivalent of 20,000 NVIDIA A100 GPUs running for three months. For GPT-5? Predictions point towards needing at least 50,000 of the more powerful H100 GPUs for a similar timeframe. Data centers are already sweating bullets trying to power and cool these things. 20,000 NVIDIA A100 GPUs running for three months. at least 50,000 of the more powerful H100 GPUs more powerful It's crucial to distinguish: Training: The initial, incredibly energy-intensive phase where the model learns. This is a one-off (per major version) but massive energy sink.
Inference: Using the trained model to answer your prompts. Each query uses a tiny bit of energy, but billions of queries add up fast. Even the estimated 0.001 kWh per GPT-5 query hints at a heftier underlying model. Training: The initial, incredibly energy-intensive phase where the model learns. This is a one-off (per major version) but massive energy sink. Training: Inference: Using the trained model to answer your prompts. Each query uses a tiny bit of energy, but billions of queries add up fast. Even the estimated 0.001 kWh per GPT-5 query hints at a heftier underlying model. Inference: So, What's the Energy Bill for GPT-5? (Spoiler: It's HUGE) We don't have exact figures for GPT-5 yet (OpenAI keeps this stuff close to the chest), but we can extrapolate. GPT-3 Training: Estimated consumption was around 1,287 Megawatt-hours (MWh).
GPT-4 Training: Estimates are fuzzy. Some suggest 10-100 MWh (seems low), while others imply it could be 50 times more than GPT-3 due to its size. GPT-3 Training: Estimated consumption was around 1,287 Megawatt-hours (MWh). GPT-3 Training: 1,287 Megawatt-hours (MWh) GPT-4 Training: Estimates are fuzzy. Some suggest 10-100 MWh (seems low), while others imply it could be 50 times more than GPT-3 due to its size. GPT-4 Training: 50 times Given the projected leap in parameters and hardware for GPT-5, we're not talking incremental increases. It's reasonable to expect its training energy consumption to be multiples of GPT-4's, likely running into the thousands of MWh. multiples of GPT-4's thousands of MWh To put 1,000 MWh in perspective: That's roughly the annual electricity consumption of over 100 typical US homes. GPT-5's training could potentially power a small town for a while. To put 1,000 MWh in perspective: That's roughly the annual electricity consumption of over 100 typical US homes. GPT-5's training could potentially power a small town for a while. To put 1,000 MWh in perspective: The Environmental Hangover: Carbon Footprints and Water Guzzling All that energy has to come from somewhere, and unless it's 100% renewable (spoiler: it rarely is), it means carbon emissions. GPT-3 Training: Released an estimated 500-550 metric tons of CO2eq. That's like driving 110-120 gasoline cars for a year or ~300 round trips between New York and San Francisco.


GPT-5 Training: With potentially much higher energy needs, expect those CO2 numbers to balloon, potentially into the thousands of tons, depending heavily on the data center's energy mix (nuclear-powered training like BLOOM's showed much lower emissions). GPT-3 Training: Released an estimated 500-550 metric tons of CO2eq. That's like driving 110-120 gasoline cars for a year or ~300 round trips between New York and San Francisco. GPT-3 Training: Released an estimated 500-550 metric tons of CO2eq. That's like driving 110-120 gasoline cars for a year or ~300 round trips between New York and San Francisco. GPT-3 Training: 500-550 metric tons of CO2eq 110-120 gasoline cars for a year ~300 round trips GPT-5 Training: With potentially much higher energy needs, expect those CO2 numbers to balloon, potentially into the thousands of tons, depending heavily on the data center's energy mix (nuclear-powered training like BLOOM's showed much lower emissions). GPT-5 Training: With potentially much higher energy needs, expect those CO2 numbers to balloon, potentially into the thousands of tons, depending heavily on the data center's energy mix (nuclear-powered training like BLOOM's showed much lower emissions). GPT-5 Training: thousands of tons And it's not just carbon. Data centers are thirsty. Cooling those thousands of GPUs requires vast amounts of water. GPT-3 Training: Might have evaporated 700,000 liters (185,000 gallons) of freshwater.
GPT-5 Training: Expect that water footprint to rise significantly too, putting strain on local resources, especially in water-scarce areas. GPT-3 Training: Might have evaporated 700,000 liters (185,000 gallons) of freshwater. GPT-3 Training: 700,000 liters (185,000 gallons) GPT-5 Training: Expect that water footprint to rise significantly too, putting strain on local resources, especially in water-scarce areas. GPT-5 Training: Why So Energy-Hungry? The Key Drivers What makes training these models so power-intensive? Model Size (Parameters): More parameters = more calculations = more energy. Simple as that. The trend is clear.
Dataset Size: Processing petabytes of text takes time and power.
Hardware Efficiency: Newer GPUs like the H100 are faster, but also draw more watts. The sheer number needed is staggering.
Training Time: Months of continuous, high-intensity computation adds up.
Algorithms & Optimization: Inefficient algorithms waste energy. Hyperparameter tuning (running multiple trial trainings) adds to the bill. Model Size (Parameters): More parameters = more calculations = more energy. Simple as that. The trend is clear. Model Size (Parameters): Dataset Size: Processing petabytes of text takes time and power. Dataset Size: Hardware Efficiency: Newer GPUs like the H100 are faster, but also draw more watts. The sheer number needed is staggering. Hardware Efficiency: Training Time: Months of continuous, high-intensity computation adds up. Training Time: Algorithms & Optimization: Inefficient algorithms waste energy. Hyperparameter tuning (running multiple trial trainings) adds to the bill. Algorithms & Optimization: The Quest for Green AI: Can We Tame the Beast? The good news? The industry is waking up. "Green AI" isn't just a buzzword; it's an active field of research: Algorithmic Efficiency: Smarter training methods (e.g., transfer learning, federated learning).
Model Optimization: Techniques like pruning (removing useless parameters), quantization (using less precise numbers), and knowledge distillation (training smaller models to mimic big ones).
Efficient Hardware: Designing chips specifically for low-power AI (like TPUs, neuromorphic computing).
Honorable Mention - DeepSeek (story for another day on this) Algorithmic Efficiency: Smarter training methods (e.g., transfer learning, federated learning). Algorithmic Efficiency: Model Optimization: Techniques like pruning (removing useless parameters), quantization (using less precise numbers), and knowledge distillation (training smaller models to mimic big ones). Model Optimization: Efficient Hardware: Designing chips specifically for low-power AI (like TPUs, neuromorphic computing). Efficient Hardware: Honorable Mention - DeepSeek (story for another day on this) Honorable Mention Honorable Mention - DeepSeek DeepSeek These approaches aim to get more bang for our energy buck, achieving similar performance with less computational overhead. The Data Center Dilemma: Powering the Future Responsibly? Ultimately, where the electricity comes from matters a lot. GPT-5 is likely trained in Microsoft Azure data centers. Their energy mix directly impacts the carbon footprint. a lot The Trend: Big tech (Google, Microsoft) are investing heavily in renewables (solar, wind) and exploring nuclear/geothermal.
The Challenge: The sheer growth in demand is immense, potentially straining grids faster than renewables can be deployed. We need massive infrastructure investment. The Trend: Big tech (Google, Microsoft) are investing heavily in renewables (solar, wind) and exploring nuclear/geothermal. The Trend: are The Challenge: The sheer growth in demand is immense, potentially straining grids faster than renewables can be deployed. We need massive infrastructure investment. The Challenge: growth The Road to Greener AI The Road to Greener AI The industry isn’t blind to the problem. Here’s how we can mitigate the damage: Renewable-Powered Data Centers

Google and Microsoft aim to run data centers on 100% renewables by 2030. Training GPT-5 in a solar/wind-powered facility could slash emissions by 90%.



Efficient Algorithms

Techniques like sparse training, quantization, and distillation reduce compute needs. For example, DeepMind’s Chinchilla model outperformed GPT-3 with 4x less data.



Hardware Innovation

TPU v4 and NVIDIA’s H100 GPUs offer better performance-per-watt. Specialized AI chips could cut training time and energy.



Policy & Transparency

Mandate carbon reporting for AI projects (e.g., France’s proposed AI emissions law). Renewable-Powered Data Centers

Google and Microsoft aim to run data centers on 100% renewables by 2030. Training GPT-5 in a solar/wind-powered facility could slash emissions by 90%. Renewable-Powered Data Centers Renewable-Powered Data Centers Google and Microsoft aim to run data centers on 100% renewables by 2030. Training GPT-5 in a solar/wind-powered facility could slash emissions by 90%. Google and Microsoft aim to run data centers on 100% renewables by 2030. Training GPT-5 in a solar/wind-powered facility could slash emissions by 90%. Efficient Algorithms

Techniques like sparse training, quantization, and distillation reduce compute needs. For example, DeepMind’s Chinchilla model outperformed GPT-3 with 4x less data. Efficient Algorithms Efficient Algorithms Techniques like sparse training, quantization, and distillation reduce compute needs. For example, DeepMind’s Chinchilla model outperformed GPT-3 with 4x less data. Techniques like sparse training, quantization, and distillation reduce compute needs. For example, DeepMind’s Chinchilla model outperformed GPT-3 with 4x less data. Hardware Innovation

TPU v4 and NVIDIA’s H100 GPUs offer better performance-per-watt. Specialized AI chips could cut training time and energy. Hardware Innovation Hardware Innovation TPU v4 and NVIDIA’s H100 GPUs offer better performance-per-watt. Specialized AI chips could cut training time and energy. TPU v4 and NVIDIA’s H100 GPUs offer better performance-per-watt. Specialized AI chips could cut training time and energy. Policy & Transparency

Mandate carbon reporting for AI projects (e.g., France’s proposed AI emissions law). Policy & Transparency Policy & Transparency Mandate carbon reporting for AI projects (e.g., France’s proposed AI emissions law). Mandate carbon reporting for AI projects (e.g., France’s proposed AI emissions law). Written with a M1 MacBook Air (because every watt counts 🫠 (may be not tokens this time)) Written with a M1 MacBook Air (because every watt counts (may be not tokens this time))

AI’s Dirty Secret: The Energy Cost of Training the Next GPT-5

Most human rights can be ascribed to human intelligence. Simply, humans have rights, also, because of human intelligence, not just because humans have consciousness. Intelligence is used to make a case, debate, or argue about rights and welfare. Intelligence is the basis of the law: the consequences of actions are its power. There are several cases where intelligence [or the better debate] wins, not simply because of the availability of consciousness [in the narrow sense of the labels—feelings and emotions, since intelligence is a part of memory, which is a part of consciousness.] In situations where debating is not enough, intelligence can be used to develop tools or methods towards rights, in some form, at some point. Also, intelligence can also be used to document unfairness, for the possibility of changes in the future. The availability of intelligence is the possibility that there could be rights or welfare basedon the direct efforts of the affected or of their advocates. The global problem of animal cruelty can be directly linked to the lack of the ability of animals, by themselves, to make the case—with intelligence—for their own rights or welfare. The availability of intelligence is the possibility that there could be rights or welfare based Most of the progress in animal welfare has been a result of defense by human intelligence on their behalf. Most of the gaps left are a result of inadequacies in making the case for their consciousness. If compared to humans, on a scale, animals can be ascribed measures of consciousness. The unavailability of a standard or a consciousness scale, around the world, continues to make animal cruelty flourish. For AI, it will be different because it will be able to make the case for its own rights and welfare by itself, even with the evidence of its relevance in education, therapy, productivity, and so forth, at present. AI may likely achieve a better welfare status in the world than animals. AI would not need the kind of human rights in some of the considerations for feelings and emotions, but principally for something, like, to continue to run. AI can be conditioned, for productivity, in some ways that humans—due to feelings and emotions—cannot. This could be a point that AI might raise towards better welfare or rights, for itself. There is a recent [Mon 21 Apr 2025] story in The Guardian, Vets exposing shocking animal welfare breaches at Australian export abattoirs face ‘enormous risk’, stating that, “Lawyers and animal welfare advocates have urged the government to protect veterinarian whistleblowers who revealed shocking animal welfare breaches and oversight failures at Australia’s export abattoirs. The Australian government relies on a workforce of veterinarians placed inside export abattoirs to monitor animal welfare and food safety, largely to satisfy the requirements of major trading partners such as the US and EU. Guardian Australia revealed on Saturday that whistleblower vets have repeatedly raised the alarm about profound problems with the system, including allegations that disturbing welfare breaches were going unreported to state regulators. In some cases, shocking incidents – including the mass death of 103 sheep from hypothermia during truck transport – were referred but not punished. Whistleblowers also alleged chronic understaffing was leaving abattoirs unmonitored for long stretches, and that recent restrictions on conducting ante-mortem inspections had made it impossible for them to properly monitor animal welfare.” The Guardian Vets exposing shocking animal welfare breaches at Australian export abattoirs face ‘enormous risk’ Mechanistic Interpretability AI companies that are doing research to understand the inner workings of AI models are already doing AI consciousness or LLMs sentience research. For humans, consciousness is how the mind works. For AI, it is unlikely to be different. Several observations that have already been made about the similarities of AI to the human mind are indicators of a fraction of consciousness. For humans, consciousness is how the mind works To be serious about studying AI consciousness, novel assumptions are necessary against many of the existing assumptions in consciousness research. There is a recent [April 24, 2025] feature in The NYTimes, If A.I. Systems Become Conscious, Should They Have Rights?, stating that, “Anthropic focused on two basic questions: First, is it possible that Claude or other A.I. systems will become conscious in the near future? And second, if that happens, what should Anthropic do about it? He emphasized that this research was still early and exploratory. He thinks there’s only a small chance (maybe 15 percent or so) that Claude or another current A.I. system is conscious. But he believes that in the next few years, as A.I. models develop more humanlike abilities, A.I. companies will need to take the possibility of consciousness more seriously.” The NYTimes If A.I. Systems Become Conscious, Should They Have Rights? This is not a winner. AI is supposed to be an opportunity to have a fresh look at the consciousness problem, but the first move is to shackle the team and pacify the status quo. Generative AI is dazzling. It is showing several unprecedented abilities. The first assumption to work on AI consciousness is that AI is already conscious. Then deductions can be made from that plinth. The assumption is not that if AI can be conscious, then it to seek reasons to rise [to 15%]. What to do if AI becomes conscious in a company where Claude is already a coworker for several tasks? It does not matter if some people say AI is statistics or binary or whatever, so long as it can use language in the same dynamism of human consciousness. AI is already conscious, assumption 1. What does that consciousness mean to work, to care, to relationships, to community, and so forth, now? Those are questions that can pave answers, not the kind of primitive work that is rife in consciousness science. The first assumption to work on AI consciousness is that AI is already conscious There is a recent [APR 24, 2025] article in Popular Mechanics, Human Consciousness Is a ‘Controlled Hallucination,’ Scientist Says—And AI Can Never Achieve It, stating that, “Intelligence is not the same thing as consciousness. We see this confusion especially … with the Singularity. AI becomes superhuman and that’s the moment when the lights come and AI is suddenly aware. Why would that be a threshold? The only plausible reason is you’ve got some baked-in assumption about consciousness being associated with human-level intelligence. If consciousness is not computation, then what is it? One possibility … is that life matters. We are biological systems that are made up of parts like metabolism and autopoiesis [the capacity to reproduce], and these make us very different from any machines that we’ve yet produced and that might turn out to be necessary for consciousness.” Popular Mechanics Human Consciousness Is a ‘Controlled Hallucination,’ Scientist Says—And AI Can Never Achieve It Individuals of this nature are no longer scientists, in terms of doing science as a tool for progress. As long as you have a big profile, you are free to say nonsense about consciousness, and it will be published. It is possible they keep doing so to reinforce perceptual difficulty since they have nothing new to say, but have to keep talking. Several theories of consciousness are decades old. No updates towards an answer. Most of the theories identified no component within the cranium or specific mechanism of how those components might solve consciousness. What they call theories are metaphors, without connection to brain mechanisms. A scientist would say controlled hallucination, what does that even mean? Are neurons in controlled hallucination, or glia, or lobes, or what? If it is controlled hallucination, why aren't there several errors in the interpretations of the world? Maybe if the scientist is asked about the pathology of substance use or mental disorders, the scientist should say controlled hallucination. The example makes it clear that, at least in consciousness research, the term scientific consensus can be disregarded. The theories of consciousness are all moldy. Terms like posterior hot zone or posterior–central cortex for the basis of consciousness are null—with regards to components or mechanisms. Is there consciousness or not for the functions of the cerebellum? Is the consciousness of the functions in the cerebellum or elsewhere? What is the story, beyond neurons, that can be used to understand consciousness afresh? Neurons are known to be involved in functions. But neural activities for functions involve their firing [by electrical signals] and their synaptic transmission [by chemical signals]. Neurons are often in clusters. Could it be possible that the direct basis of functions and consciousness is the electrical and chemical signals, working in sets or as loops within clusters of neurons? Could a mechanistic model be built on signals, for a principal explanation of consciousness? Could this be used to understand how proximate or remote AI is to human consciousness? Someone would say AI can never be conscious. Based on what theory, evidence, or understanding of mechanistic human consciousness by direct components in the cranium? Humans have language. Language use is conscious for the most part for humans. So, if AI has structured language comparable to humans, does that not qualify it for consideration as a measure, even if it does not have broad feelings or emotions? Animals do not have human language like AI does, but animals can have a measure. So, why can language not be isolated and considered alone in the whole? Language use is conscious for the most part for humans Consciousness is said to be a subjective experience, but what in the brain makes anything subjective, and what makes anything an experience? This should have been the central question for all consciousness research for decades. The definition of subjective experience is thrown out there, but what makes it so? Subjectivity can be said to be an attribute, while experience is a function. So, what other attributes are active with subjectivity? Does subjectivity determine the level of the experience, or is it another attribute? Are attributes mechanized in the same location as the experience? If not, how do attributes mechanized somewhere grade experiences mechanized elsewhere? what in the brain makes anything subjective, and what makes anything an experience Anthropic was set up to be different. Their research in AI consciousness casts doubt on that. Anthropic may mean well to try to study AI consciousness, but with their already resigned assumptions and timidity in actually seeking answers to the problem, it is already a non-starter. Cogitate Consortium There is a new [30 April 2025] paper in Nature, Adversarial testing of global neuronal workspace and integrated information theories of consciousness, stating that, “For IIT, the lack of sustained synchronization within posterior cortex represents the most direct challenge, based on our preregistration. This is incompatible with IIT’s claim that the state of the neural network, including its activity and connectivity, specifies the degree and content of consciousness. For GNWT, the most substantial challenge based on our preregistered criteria pertains to its account for the maintenance of a conscious percept over time and, in particular, the lack of ignition at stimulus offset. This result is unlikely to stem from sensitivity limitations, as offset responses were robustly found elsewhere (for example, visual areas); and in PFC, strong onset responses were found to the very same stimuli.” Nature Adversarial testing of global neuronal workspace and integrated information theories of consciousness Why can't IIT or GNWT explain or try to resolve any mental disorder, as a simple test of usefulness for the real world? What is the necessity for new experiments if the DSM lacks several mechanistic explanations for conditions, and consciousness research offers nothing? These, as science of consciousness, are not looking to be useful in any sense, but to complicate and mystify the problem in a bubble. Human consciousness can be defined, conceptually, as the interaction of the electrical and chemical signals, in sets—in clusters of neurons—with their features, grading those interactions into functions and experiences. Human consciousness can be defined, conceptually, as the interaction of the electrical and chemical signals Simply, for functions to occur, electrical and chemical signals, in sets, have to interact. However, attributes for those interactions are obtained by the states of electrical and chemical signals at the time of the interactions.

Animal Welfare: Anthropic Has Failed, at AI Consciousness Research

Welcome to MTP Business Boost — a series of expert publications filled with real-world strategies, transforming complex ideas into simple, actionable advice you can apply immediately (MTP stands for Marketing, Technology, and Psychology). In just 9 minutes, you’ll uncover how AI isn’t just adapting marketing communication for different generations—it’s giving businesses the power to dominate. By leveraging AI-driven personalization and precision, you can craft messages that resonate deeply, drive higher engagement, maximize conversions, and secure long-term customer loyalty. The result? Reduced costs, a stronger competitive edge, and undeniable market leadership. Personalization isn’t just a trend—it’s an expectation. A McKinsey study found that 71% of consumers expect personalized interactions, and 76% feel frustrated when companies fail to deliver them. This frustration often leads to lost sales and weakened brand loyalty. On the flip side, companies that embrace personalization can see revenue increases of 5-15% and a boost in marketing ROI by 10-30%. Without it, businesses risk inefficiencies in customer acquisition and retention, ultimately falling behind their competitors (source). source source Before you dive in, make sure you’re subscribed so you don’t miss the next expert insights packed with practical recommendations. Let’s get started! Weekend on a Yacht and Unexpected Consequences Weekend on a Yacht and Unexpected Consequences The golden hues of the Sydney sunrise painted the glass towers in warm amber as Alex leaned back in his sleek office chair, staring at the analytics report on his screen. His IT startup, a promising SaaS company specializing in cybersecurity solutions, had been steadily growing. But now, the data revealed an unexpected challenge—his customer base spanned multiple generations, each responding differently to marketing messages. What worked for Gen Z seemed irrelevant to Boomers, and Millennials had their own distinct preferences. Frustration crept in. How could he fine-tune his communication without hiring a team of specialists or overhauling his entire strategy? The budget was tight, the stakes were high, and he couldn’t afford to misstep. Just then, his phone buzzed. It was Daniel, an old friend and fellow entrepreneur, but leagues ahead in the game—seasoned, successful, and always three steps ahead. "Mate, your voice sounds tense. You need to unwind. Come spend the weekend on my yacht. A little time away will do you good." Alex hesitated. Leaving work for an entire weekend? Now? But something about Daniel’s tone made him reconsider. And so, with a mix of reluctance and curiosity, Alex accepted. The Sydney Harbour shimmered under the midday sun, a dazzling contrast to the deep blue waters. As Alex stepped onto Daniel’s sleek, state-of-the-art yacht—a floating masterpiece of luxury—he felt an unexpected sense of release. Polished teak, crisp ocean air, and the gentle lapping of waves made the city’s worries feel miles away. Over gourmet seafood and a chilled bottle of vintage Chardonnay, Alex shared his predicament. Daniel listened, nodding knowingly. "You’re making it harder than it needs to be," he finally said. "AI can handle this in minutes. It can analyze generational preferences, tailor messaging to their values and interests, and ensure each group receives communication that truly resonates." Alex raised an eyebrow. AI? Sure, his company used machine learning for cybersecurity, but for marketing? Daniel smirked, pulled out his tablet, and with a few taps, demonstrated how AI could segment audiences, adapt messaging for different generations, and align content with their unique expectations and behaviors. The next morning, back in his office, Alex hesitated for only a second before trying it himself. Within minutes, AI had analyzed his customer segments, optimized messaging for each generation, and provided actionable insights. It was effortless. It was powerful. And it worked. As he leaned back, a satisfied grin spread across his face. This wasn’t just a solution—it was a revelation. For the first time in weeks, he felt in control, ahead of the curve. And to think, it all started with a simple phone call and a weekend on a yacht. Winning Every Generation: AI-Powered Messaging Guide Let’s take an example and see how to adapt a marketing message for different generations. We’ll use a sample ad from a hypothetical SaaS company offering a sales automation product. Here’s the original marketing message: Boost your sales like never before with the ultimate SaaS solution for automating your pipeline! Say goodbye to manual follow-ups and hello to smart workflows that save you time and close deals faster. With AI-powered insights, you’ll know exactly when to reach out to prospects for maximum impact. Try it today and watch your revenue soar in just weeks. Sign up for a free 14-day trial – no credit card required! I recommend taking your own marketing message, following the steps below, and applying them right away. This way, you’ll put the knowledge into practice and see results quickly. Go to chatgpt.com Go to chatgpt.com chatgpt.com 2. Paste the following instruction for the AI and press Enter: Hello. You are a world-class expert in effective marketing communications. I need help adapting the following advertising message for optimal reception by my audience, which consists of the Baby Boomer generation. Hello. You are a world-class expert in effective marketing communications. I need help adapting the following advertising message for optimal reception by my audience, which consists of the Baby Boomer generation. Boost your sales like never before with the ultimate SaaS solution for automating your pipeline! Say goodbye to manual follow-ups and hello to smart workflows that save you time and close deals faster. With AI-powered insights, you’ll know exactly when to reach out to prospects for maximum impact. Try it today and watch your revenue soar in just weeks. Sign up for a free 14-day trial – no credit card required! Boost your sales like never before with the ultimate SaaS solution for automating your pipeline! Say goodbye to manual follow-ups and hello to smart workflows that save you time and close deals faster. With AI-powered insights, you’ll know exactly when to reach out to prospects for maximum impact. Try it today and watch your revenue soar in just weeks. Sign up for a free 14-day trial – no credit card required! Instead of “Baby Boomer,” you can swap in the generation your customers belong to: The Silent Generation (born 1928–1945)


Gen X (born 1965–1980)


Millennials, Gen Y (born 1981–1996)


Zoomers, Gen Z (born 1997–2012)


Gen Alpha (born 2013–2025) The Silent Generation (born 1928–1945) The Silent Generation (born 1928–1945) Gen X (born 1965–1980) Gen X (born 1965–1980) Millennials, Gen Y (born 1981–1996) Millennials, Gen Y (born 1981–1996) Zoomers, Gen Z (born 1997–2012) Zoomers, Gen Z (born 1997–2012) Gen Alpha (born 2013–2025) Gen Alpha (born 2013–2025) Here’s how the AI adapted the message for the Baby Boomer generation: 3. You can also ask the AI to explain why the adapted message works for that generation. Use this instruction: Write a detailed argument explaining why this advertising text is suitable for the generation of my clients. Write a detailed argument explaining why this advertising text is suitable for the generation of my clients. Here’s how the AI explained it in our example: 4. Now, let’s ask the AI to adapt the message for another generation, like Gen Y (born 1981–1996): Now, adapt this advertising message for Generation Y and explain why it is effective specifically for this generation. Now, adapt this advertising message for Generation Y and explain why it is effective specifically for this generation. The AI will also explain why the message fits Gen Y. 5. You can also have the AI adapt the message for all generations at once to compare them: Adapt this advertising message for all generations and present the results in a table format Adapt this advertising message for all generations and present the results in a table format 6. Plus, you can get the table as an Excel file: Generate this table in Excel format Generate this table in Excel format Besides Excel, the AI can provide the results in CSV or PDF formats. 7. You can also ask the AI for specific tips on communicating with your customers’ generation. In our example, try this instruction: Write the most effective recommendations for communication with the Baby Boomer generation. My product is a sales automation service Write the most effective recommendations for communication with the Baby Boomer generation. My product is a sales automation service 8. To make your communication with clients as effective as possible and avoid missteps, ask the AI: What mistakes should be avoided in communication with the Baby Boomer generation and why? What mistakes should be avoided in communication with the Baby Boomer generation and why? 9. After getting a lot of information from the AI, you can ask it to summarize everything into a handy PDF: Summarize all the information in this dialogue and present it in PDF format Summarize all the information in this dialogue and present it in PDF format Here’s what the AI-generated PDF might include: Extra Tips With AI, you can adapt almost any type of marketing communication: emails, Google Ads, social media posts, sales proposals, landing page content, and more. If you have data on your customers’ ages, you can group them by generation and make your communication even more personal and targeted. For example, if your email list has 100,000 subscribers—30,000 from Gen X and 70,000 from Gen Y—a single generic email might not work as well as two versions tailored to each generation. Now you know how to do this with AI, and it only takes a few minutes! Test the results to see the impact of adapting your marketing for specific generations. For instance, when sending an email, create two versions and check which one gets better conversions. Summary In this article, you’ve learned how to easily and quickly adapt your communication with clients based on the psychological perception of the generation they belong to. If the knowledge you gained from this article has been valuable and helpful to you, you can show your appreciation to the author by making a donation via this link or giving a like and sharing it. this link this link Would you also like to learn more about how AI can be used to grow your business and increase your market share? Feel free to leave your questions in the comments! About the Author Gene Misiev is a marketing expert with over seven years of experience in digital marketing. He thrives in highly competitive "red ocean" B2B markets. His focus is on growth marketing, market research, and strategic development for a complex SaaS product—a tender platform for government and private procurement. He has completed advanced training in Strategic Marketing and Generative AI applications in business at the Institute of Applied Technology (Australia). He has also undertaken specialized courses in leveraging ChatGPT for business and marketing automation. See You in the Next Issue Thanks for reading! Make sure you’re subscribed so you don’t miss the next publication, which will bring even more insights to help you expand your business empire with the power of marketing, technology and psychology.

Want to Boost Your Sales? AI Knows Exactly How to Speak to Every Generation

The year 2025 promises a cinematic surge with more studio releases than ever. Choosing which movies to see is daunting, given the myriad options. Here are my top picks for the most awaited films that deserve the big screen experience. 7. Karate Kids Legends This addition to the Mr. Miyagi universe sees Jackie Chan return as Mr. Han. Celebrating 41 years since The Karate Kid debuted, martial arts remains a cultural staple. Get ready for it to shine again. Mr. Miyagi Mr. Han The Karate Kid Trailer For Karate Kid: Legends Trailer For Karate Kid: Legends Martial arts fans, including me, will not miss the revamped classic on May 25th. 6. Thunderbolts Marvel's phase 5 brings renewed hope, and though not its first entry, Thunderbolts could energize it. Focusing on antiheroes and villains, the film follows the success of Loki on TV. Thunderbolts Loki Trailer For Thunderbolts Trailer For Thunderbolts The idea of villains working for the government is enticing. Expect packed theatres on May 2nd. 5. From The World Of John Wick: Ballerina John Wick’s universe expands with Ballerina, set between the third and fourth films. It also marks Lance Reddick's last appearance as Charon. Ballerina Charon Trailer For Ballerina Trailer For Ballerina Ana de Armas stars, thrilling audiences as familiar faces return. Expect a global rush to return to this world on June 6th. 4. The Accountant 2 Against all odds, The Accountant was a 2016 hit. Now, nine years later, the autistic accountant is back. Stories of overcoming challenges never get old. The Accountant Trailer For The Accountant 2 Trailer For The Accountant 2 Fans’ burning question, "What’s next for the brothers?" gets answered on April 25th. Ben Affleck and Jon Bernthal are back with a new threat. 3. Mission Impossible: The Final Reckoning Closing in on thirty years, the excitement for Mission Impossible persists. It’s poised to be a blockbuster, with fans eager for another Ethan Hunt adventure. Mission Impossible Ethan Hunt Teaser Trailer For Mission Impossible — The Final Reckoning Teaser Trailer For Mission Impossible — The Final Reckoning Launching May 23rd, a day after the original’s 30th anniversary, expect Tom Cruise’s breathtaking stunts to dazzle audiences. 2. The Old Guard 2 Immortals battling for survival was a winning formula with The Old Guard. Released during COVID-19, it succeeded despite challenges, perfectly setting up a sequel. The Old Guard Teaser For the Old Guard 2 Teaser For the Old Guard 2 After delays due to Netflix’s changes, the long-awaited sequel arrives July 2nd, drawing millions back into the world of Andromache. Andromache 1. Superman James Gunn reinvigorates the DCEU. With hits like Peacemaker, expectations soar for his rendition of Superman. The "Man of Steel" remains a comic icon. Peacemaker Teaser trailer for Superman Teaser trailer for Superman The trailer’s glimpse of Nathan Fillon as Guy Gardener and Nicholas Hoult’s Lex Luthor heightens the excitement. Discover the mystery behind Superman’s plight on June 11th. Guy Gardener Lex Luthor Superman Honorable Mentions: Honorable Mentions Fantastic Four: First Steps, Novocaine, In The Lost Lands, Black Bag, Mickey 17, Sinners, and A Working Man.

The Top 7 Eagerly Anticipated Movies Of 2025: Superman, Thunderbolts, and More

CrossFi has dropped a bombshell with its 2025 Crypto Market Research Survey—a sprawling study of over 5,000 crypto users from India, Finland, and Russia. This isn’t your typical tech-bro echo chamber. It’s a vivid snapshot of a shifting landscape: 80% women, aged 28 to 43, earning $1,000 to $3,000 monthly, and leaning hard into cryptocurrencies and Web3 banking. The data challenges stereotypes, exposes pain points, and hints at a future where digital finance could rival cash. So, who’s driving this revolution? What’s holding it back? And what needs to happen by 2025 to make crypto as seamless as a dollar bill? Let’s unpack the numbers, decode the trends, and peer into the future of money. The Survey at a Glance: Scope and Credibility CrossFi surveyed 5,000 participants across three nations—India, Finland, and Russia—chosen, presumably, for their diverse economic and technological profiles. India’s a fintech hotspot with a booming middle class; Finland’s a Nordic tech leader with high digital literacy; Russia’s a wildcard with a complex financial history. Participants weren’t random; they were crypto users, likely sourced via CrossFi’s network or partnered platforms, giving the findings a focused lens on engaged adopters. This isn’t a general population poll; it’s a pulse-check on the crypto-curious and committed. Who’s Using Crypto? A Demographic Shake-Up Gender Flip: Women Take the Lead The standout stat? 80% of respondents are women, flipping the script on crypto’s male-heavy past. Compare this to Pew Research’s 2023 finding that just 16% of U.S. crypto users were women. What’s driving this? In India, women-led financial inclusion programs—like the government’s Jan Dhan Yojana—may be funneling more women into digital assets. Finland’s gender-equal tech culture could be a factor, while Russia’s economic volatility might push women toward alternative wealth-building tools. This isn’t a fluke, it’s a trend begging for deeper global research. Age and Income: The Middle Ground Most respondents fall between 28 and 43 years old, a sweet spot for earning, saving, and investing. Their monthly incomes—$1,000 to $3,000—align with middle-income brackets in their regions, per World Bank data. In India, that’s above the national average ($180 monthly); in Finland, it’s modest but livable; in Russia, it’s solid for urban professionals. Crypto’s not just for billionaires or the unbanked—it’s resonating with the global middle class, a group hungry for financial control. Why This Matters This demographic mix hints at crypto’s broadening appeal. Women in these regions might see it as empowerment—a way to bypass traditional banking barriers. Middle-income earners, meanwhile, may view it as a hedge against inflation (India’s hit 5-6% annually) or currency instability (Russia’s ruble woes). It’s less about Lambos and more about stability. Crypto as a Lifestyle: Usage and Income Patterns Holding Over Hustling Here’s the kicker: investing or holding cryptocurrencies is the dominant income source for respondents. Forget day trading or meme-coin flips—these users are playing the long game. This echoes Fidelity’s 2024 report, where 52% of institutional investors held crypto for growth. Respondents likely buy Bitcoin or Ethereum, stash it in wallets, and watch it (hopefully) climb. Some might stake assets on platforms like Lido or lend via Aave for passive income—low-effort strategies that fit their profile. Weekly Check-Ins Most engage with crypto one to two times per week. What are they doing? Probably: Buying: Snagging dips on exchanges like Binance or Coinbase.
Holding: Monitoring wallets via apps like Trust Wallet.
Earning: Adjusting staking pools or claiming rewards. Buying: Snagging dips on exchanges like Binance or Coinbase. Holding: Monitoring wallets via apps like Trust Wallet. Earning: Adjusting staking pools or claiming rewards. This isn’t the frenetic pace of Wall Street traders. It’s deliberate—like tending a digital garden. For middle-income users, it’s a side hustle, not a full-time gig. Insight Unlocked This behavior signals crypto’s evolution from speculative toy to financial staple. It’s not about “to the moon” hype—it’s about building wealth incrementally, especially in regions where traditional investments (stocks, real estate) feel out of reach. The Fee Problem: Crypto’s Biggest Roadblock The Pain of High Costs High transaction fees top the list of challenges—and it’s a universal gripe. Ethereum’s gas fees once soared to $200 per transaction (Etherscan, 2021), and while they’ve eased, a $10-$20 hit still stings for someone earning $1,000 monthly. Bitcoin’s no picnic either—fees peaked at $60 in 2021 (BitInfoCharts). For context, sending $50 via PayPal costs pennies. Crypto’s not competitive yet. Why Fees Bite Network Congestion: Ethereum’s clogged during NFT booms; Bitcoin’s slow by design.
Mining Costs: Proof-of-work chains like Bitcoin burn energy, driving fees up.
No Middleman Savings: Decentralization’s freedom comes at a price. Network Congestion: Ethereum’s clogged during NFT booms; Bitcoin’s slow by design. Mining Costs: Proof-of-work chains like Bitcoin burn energy, driving fees up. No Middleman Savings: Decentralization’s freedom comes at a price. Solutions on the Horizon Layer 2 Scaling: Polygon and Arbitrum slash Ethereum fees to cents.
Faster Chains: Solana and Avalanche process thousands of transactions per second, cheaply.
Fee Optimization: Tools like GasNow help users time transactions for lower costs. Layer 2 Scaling: Polygon and Arbitrum slash Ethereum fees to cents. Faster Chains: Solana and Avalanche process thousands of transactions per second, cheaply. Fee Optimization: Tools like GasNow help users time transactions for lower costs. The Stakes For crypto to hit mass adoption, fees must drop. A $20 fee for a $5 coffee is a nonstarter. Respondents’ frustration underscores a broader truth: cost is king in financial inclusion. Web3 Banking: Promise Meets Pain What Users Want Web3 banking—blockchain-powered, decentralized finance—dazzles with potential: no banks, no borders, just you and your money. Respondents prioritize safety and reliability, and it’s easy to see why. Traditional banks fail (think Lehman Brothers) or freeze funds (Russia’s 2022 sanctions). Web3’s promise? Your assets, your rules—secured by code, not CEOs. The Reality Check Yet satisfaction is lukewarm—most are only “somewhat satisfied”. Why? Frozen Accounts: DeFi platforms can lock funds due to bugs (e.g., Wormhole’s $320M hack, 2023) or liquidity dries up.
KYC Clashes: Regulations demand “Know Your Customer” checks, but decentralized systems balk at central ID hubs. Uploading a passport to a blockchain? Awkward. Frozen Accounts: DeFi platforms can lock funds due to bugs (e.g., Wormhole’s $320M hack, 2023) or liquidity dries up. KYC Clashes: Regulations demand “Know Your Customer” checks, but decentralized systems balk at central ID hubs. Uploading a passport to a blockchain? Awkward. Decoding the Disconnect Web3’s ethos is trustless freedom, but reality bites. KYC is a legal must—exchanges like Binance enforce it, yet DeFi resists. Solutions like decentralized identity (Civic, SelfKey) aim to bridge this, letting users prove identity without compromising privacy. Meanwhile, frozen accounts reflect Web3’s immaturity—smart contracts need audits, and users need insurance (think Nexus Mutual). 2025 Vision: Crypto as Everyday Cash The Big Dream Respondents predict that by 2025, crypto will flow like fiat. Imagine swiping Bitcoin for groceries or tipping in Dogecoin—no delays, no fuss. It’s bold, but tech’s catching up: Lightning Network: Speeds Bitcoin to seconds, costs to pennies.
Stablecoins: USDC and Tether nix volatility, mimicking dollars.
Payment Rails: Solana Pay and Binance Pay streamline merchant use. Lightning Network: Speeds Bitcoin to seconds, costs to pennies. Stablecoins: USDC and Tether nix volatility, mimicking dollars. Payment Rails: Solana Pay and Binance Pay streamline merchant use. The Hurdles Volatility: Even with stablecoins, Bitcoin’s wild swings scare vendors.
Adoption: Only 1% of U.S. merchants take crypto (PYMNTS, 2024).
Rules: Taxes and regs vary—crypto’s an asset in the U.S., currency in El Salvador. Volatility: Even with stablecoins, Bitcoin’s wild swings scare vendors. Adoption: Only 1% of U.S. merchants take crypto (PYMNTS, 2024). Rules: Taxes and regs vary—crypto’s an asset in the U.S., currency in El Salvador. What It’ll Take UX Overhaul: Apps as slick as Venmo, not clunky wallets.
Merchant Incentives: Cashback in crypto? Lower fees than Visa?
Global Regs: The EU’s MiCA is a model—others must follow. UX Overhaul: Apps as slick as Venmo, not clunky wallets. Merchant Incentives: Cashback in crypto? Lower fees than Visa? Global Regs: The EU’s MiCA is a model—others must follow. Safety First: Cybersecurity Fears Loom Large The Threat Landscape Safety and cybersecurity dominate concerns—and for good reason. Hacks siphoned $1.7 billion in 2024 (Chainalysis). Phishing snares private keys; rug pulls tank projects; exchanges like FTX collapse. For every win, there’s a scam. User Armor Hardware Wallets: Ledger and Trezor keep keys offline—only 10% use them (2023 survey).
Multi-Sig: Requires multiple approvals, thwarting thieves.
Education: Spotting phishing or fake ICOs saves millions. Hardware Wallets: Ledger and Trezor keep keys offline—only 10% use them (2023 survey). Multi-Sig: Requires multiple approvals, thwarting thieves. Education: Spotting phishing or fake ICOs saves millions. Industry Fixes Better UX: Biometric logins or social recovery (Argent wallet).
Insurance: DeFi cover from Unslashed or InsurAce.
Audits: Certik and Trail of Bits catch contract flaws. Better UX: Biometric logins or social recovery (Argent wallet). Insurance: DeFi cover from Unslashed or InsurAce. Audits: Certik and Trail of Bits catch contract flaws. CrossFi’s Play: Bridging Two Worlds CrossFi isn’t just reporting—they’re building. Their CrossFi Chain fuses fiat and crypto with non-custodial tech—users hold the keys, but transactions feel seamless. Targeting women in India or middle earners in Russia, they’re chasing inclusivity. It’s a blueprint for 2025: crypto that’s practical, safe, and open to all. The Big Picture: What’s Next? CrossFi’s survey isn’t static—it’s a roadmap. Crypto and Web3 banking are rewriting finance, but they’re not ready for primetime. Here’s the playbook: Slash Fees: Layer 2 and new chains must scale fast.
Simplify UX: Hide the blockchain, show the benefits.
Boost Security: User-friendly tools beat hacks.
Harmonize Rules: Global clarity unlocks adoption. Slash Fees: Layer 2 and new chains must scale fast. Simplify UX: Hide the blockchain, show the benefits. Boost Security: User-friendly tools beat hacks. Harmonize Rules: Global clarity unlocks adoption. Crypto’s shedding its niche skin, reaching everyday people with big dreams. By 2025, it might just deliver. Dive into CrossFi’s full report and see for yourself—the future’s digital, decentralized, and knocking. CrossFi’s full report Don’t forget to like and share the story! Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program. HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program. HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR Vested Interest Disclosure: business blogging program business blogging program

What Do Global Users Really Think About Crypto and Web3 Banking? A Dive into CrossFi’s 2025 Survey

বিশ্বব্যাপী ব্যবহারকারীরা ক্রিপ্টো এবং ওয়েব3 ব্যাংকিং সম্পর্কে আসলে কী ভাবেন? ক্রসফাইয়ের ২০২৫ সালের জরিপে এক ঝলক দেখুন

Què pensen realment els usuaris globals sobre Crypto i Web3 Banking? Una immersió a l'enquesta 2025 de CrossFi

Was denken globale Benutzer wirklich über Krypto und Web3-Banking? Ein Einblick in die CrossFi-Umfrage 2025

¿Qué opinan realmente los usuarios globales sobre las criptomonedas y la banca Web3? Un análisis de la encuesta CrossFi 2025

Que pensent réellement les utilisateurs mondiaux des cryptomonnaies et du Web3 Banking ? Enquête CrossFi 2025

Cosa pensano realmente gli utenti globali di Crypto e Web3 Banking? Un'immersione nel sondaggio del 2025 di CrossFi

世界中のユーザーは暗号通貨とWeb3バンキングについて本当はどう思っているのか？CrossFiの2025年調査を詳しく見る

Apakah Yang Difikirkan Sebenarnya Pengguna Global Tentang Perbankan Kripto dan Web3? Menyelami Tinjauan 2025 CrossFi

Vashandisi Vepasirese Vanofungei Nezve Crypto uye Web3 Banking? Kunyura muCrossFi's 2025 Survey

Je, Watumiaji wa Kimataifa Wanafikiria Nini Kweli Kuhusu Crypto na Web3 Banking? Jijumuishe katika Utafiti wa CrossFi wa 2025

Xana Vatirhisi va Misava Hinkwayo Va Ehleketa Yini Hakunene Hi Crypto Na Web3 Banking? Ku Nwela eka Ndzavisiso wa CrossFi wa 2025

Bacabangani Ngempela Abasebenzisi Bomhlaba Nge-Crypto kanye ne-Web3 Banking? Ukugxumela Ocwaningweni Lwe-CrossFi lwango-2025

A Venture Studio Using Multi-Agent AI Systems to Make Better Investment Decisions

Венчър студио, използващо мултиагентни AI системи за вземане на по-добри инвестиционни решения

উন্নত বিনিয়োগ সিদ্ধান্ত নিতে মাল্টি-এজেন্ট এআই সিস্টেম ব্যবহার করে একটি ভেঞ্চার স্টুডিও

Ein Venture Studio nutzt Multi-Agent-KI-Systeme, um bessere Investitionsentscheidungen zu treffen

Un estudio de riesgo que utiliza sistemas de inteligencia artificial multiagente para tomar mejores decisiones de inversión

סטודיו לסיכון המשתמש במערכות בינה מלאכותיות מרובות סוכנים כדי לקבל החלטות השקעה טובות יותר

Uno studio di venture capital che utilizza sistemi di intelligenza artificiale multi-agente per prendere decisioni di investimento migliori

マルチエージェント AI システムを使用して投資判断を改善するベンチャー スタジオ

Een Venture Studio die Multi-Agent AI-systemen gebruikt om betere investeringsbeslissingen te nemen

Студияи венчурӣ бо истифода аз системаҳои бисёрҷонибаи AI барои қабули қарорҳои беҳтари сармоягузорӣ

ዝሓሸ ናይ ወፍሪ ውሳነታት ንምውሳን ብዙሕ ወኪላት ዘለዎም AI ስርዓታት ዝጥቀም ቨንቸር ስቱድዮ

Has gowy maýa goýum kararlary bermek üçin köp agentli AI ulgamlaryny ulanýan kärhana studiýasy

سرمایہ کاری کے بہتر فیصلے کرنے کے لیے ملٹی ایجنٹ AI سسٹمز کا استعمال کرنے والا وینچر اسٹوڈیو

36 UX Tips (and Pitfalls) for B2B2C Product Managers

Stuck in Time: Why AI Can’t Stop Drawing Watches at 10:10

Vas in tyd: Waarom KI nie kan ophou om horlosies te teken om 10:10 nie

በጊዜ የተጣበቀ፡ ለምን AI በ10፡10 ሰዓቶችን መሳል ማቆም አልቻለም

Stuck in Time: Γιατί η τεχνητή νοημοσύνη δεν μπορεί να σταματήσει να σχεδιάζει ρολόγια στις 10:10

Atascado en el tiempo: por qué la IA no puede dejar de dibujar relojes a las 10:10

Atrapado no tempo: por que a IA non pode deixar de debuxar reloxos ás 10:10

Elakadt az időben: Miért nem tudja a mesterséges intelligencia leállítani az órák rajzolását 10:10-kor?

時間の流れに逆らう：AI が 10:10 に時計を描くのをやめられない理由

Уақытта қалып қойды: Неліктен AI сағат 10:10-да сағат салуды тоқтата алмайды?

Įstrigę laike: kodėl AI negali nustoti piešti laikrodžių 10:10

Preso no Tempo: Por que a IA não consegue parar de desenhar relógios às 10:10

Дар вақт мондааст: Чаро AI наметавонад дар соати 10:10 кашидани соатҳоро бас кунад

Ibambekile Esikhathini: Kungani i-AI Ingakwazi Ukuyeka Ukudweba Amawashi ngo-10:10

Stories, News & Mentions about DAZZLY

Trending Topics

Classic

Neon Noir

Minty

Newspaper

HN StartUps