AI agents will eventually replace manual banking interfaces
βLogical conclusion is people are not going to be logging into banking websites. They're going to have their AI personal assistant, the same one that does the hair appointment for them, get their partner a gift. It'll also say, 'Hey, you received a bill for this much in your email, they applied the 20% discount they promised, and I've already scheduled it to be paid in ten days.'β
AI coding hit a massive inflection point in late 2025 - The transition from simple autocomplete to autonomous agents allows developers to build complex software entirely from their phones, fundamentally shifting the speed and nature of creation.
βNovember 2025 was the inflection point when AI coding agents crossed from 'mostly works' to 'actually works.'β
Anthropic Mythos leads the new AI cybersecurity race
βMicrosoft is actually integrating that, but it's just a preview. It's directly inside of their secure coding framework, and the idea is that Clawed gets used for threat detection, vulnerability scanning, and incident response inside of Microsoft's DevTooling. And at the same time, OpenAI has been briefing US federal agencies, state governments, and also the Five Eyes intelligence partners on a version of their model called GPT 5.5 Cyber.β
AI agents will eventually replace manual banking interfaces
βLogical conclusion is people are not going to be logging into banking websites. They're going to have their AI personal assistant, the same one that does the hair appointment for them, get their partner a gift. It'll also say, 'Hey, you received a bill for this much in your email, they applied the 20% discount they promised, and I've already scheduled it to be paid in ten days.'β
Asia's super app ecosystem accelerates agentic AI adoption
βthe super app ecosystem nature of Asia. So if you look at China, which has got a large number of super apps, if you look at Korea, which has got a super app, if you look at Indonesia or India with their own versions of super apps, then these are actually driving very significant consumer behavior given the nature of demographics of Asia. The second thing is ultimately AI comes down to, and both Michael and Lareina spoke about this, the chips and the semiconductor stack. Majority of the stack is being built in Asia.β
Anthropic valuation surpasses OpenAI on the secondary market
βOn the secondaries, OpenAI's valuation, what people are buying the shares for right now, it's kind of interesting because it's like an indicator of kind of where the stock will go sometimes. What OpenAI's shares are selling for right now is $850 billion. Anthropic's shares, $1 trillion. So Anthropic is being priced and valued at higher possible, you know, possible part of that is maybe not even where they are today, but where people can see them going in the future.β
Agentic banking will accelerate the disruption of incumbents
βFintech is only three to five percent of the market for financial services, and we've been talking for a long time about how banks are going to have to wisen up because startups are coming to eat their lunch. But the rate of that flip now with agents opening bank accounts is making traditional banking look like primitive technology. The next two to three years will be unlike anything we've seen.β
GPT 5.5 pricing introduces a high-tier pro subscription
βThe pricing, I think, is also where this gets a lot, gets very interesting. The standard GPT-55 is $5 per million input tokens and $30 per million output tokens. That doubles GPT-54 on paper, but OpenAI is claiming that the model uses tokens more efficiently, so the real world cost per task should be roughly flat or better. GPT 5.5 Pro is $30 in and $180 out, which is very expensive.β
Prompt injection remains an unsolved and catastrophic security risk - The combination of AI autonomy, data access, and the normalization of technical deviance creates a lethal trifecta that could lead to a major industry disaster.
βPrompt injection is an unsolved security problem and the 'lethal trifecta' that will likely lead to an AI Challenger disaster.β
Anthropic valuation surpasses OpenAI on the secondary market
βOn the secondaries, OpenAI's valuation, what people are buying the shares for right now, it's kind of interesting because it's like an indicator of kind of where the stock will go sometimes. What OpenAI's shares are selling for right now is $850 billion. Anthropic's shares, $1 trillion. So Anthropic is being priced and valued at higher possible, you know, possible part of that is maybe not even where they are today, but where people can see them going in the future.β
Meow launched the first bank accounts for AI agents
βWe became the first to allow an AI agentβso your favorite LLM, be it Claude, Gemini, or ChatGPTβto open an actual bank account for you. This is opening up a ton of doors because indie hackers want no friction. They want to be able to spin up a business, open a bank account, and accept payments all in one prompt.β
Engineers now review AI-generated code instead of writing
βOur engineers are not really writing code by hand as much anymore, if at all. They're reviewing code by hand and have agents to help them. We've invested a lot in our infrastructure to make it AI friendly, allowing code to be written, reviewed, and automatically triaged if there's an alert. It is absolutely bananas what's happened in the past year and a half.β
Meow launched the first bank accounts for AI agents
βWe became the first to allow an AI agentβso your favorite LLM, be it Claude, Gemini, or ChatGPTβto open an actual bank account for you. This is opening up a ton of doors because indie hackers want no friction. They want to be able to spin up a business, open a bank account, and accept payments all in one prompt.β
βThe model itself, basically OpenAI's positioning on this is that it is quote unquote a fully agentic model, meaning that it's designed to complete all these multi-step computer tasks with minimal human direction. They specifically highlighted five different categories, analyzing data, writing and debugging code, operating software directly, researching online, creating documents and spreadsheets autonomously.β
Live translation breaks cultural barriers across Asia
βthis week, I spent doing store visits and distribution. And part of the thing which I found very useful is that, you know, it makes it very easy to connect with people. The live translation capabilities, like if I do this a decade ago, then I would always be accompanied with the translator, now I'm just accompanied with my phone with super cool translation abilities. And I'm able to connect with the nations who don't speak English. And I'm able to understand essentially how the business process is running, with instructions which are all written in Bahasa, as same as in India.β
Software development is moving toward a dark factory model - We are entering a paradigm where AI handles the entire lifecycle of code creation, review, and QA, producing software at a scale that humans can no longer manually audit.
βThe next leap is the 'dark factory' pattern where nobody writes or reviews code and AI does its own QA.β
Engineers now review AI-generated code instead of writing
βOur engineers are not really writing code by hand as much anymore, if at all. They're reviewing code by hand and have agents to help them. We've invested a lot in our infrastructure to make it AI friendly, allowing code to be written, reviewed, and automatically triaged if there's an alert. It is absolutely bananas what's happened in the past year and a half.β
Financial infrastructure must remain agnostic across AI models
βWe went heavy with the Claude branding because they seem to be the far and away winner right now, especially for businesses, but we built it to be agnostic behind the scenes. You're exactly right; even though our branding is currently Claude-oriented, it is agnostic. I would be terrified to just be a pure software company right now.β
Agents create a 1% error that compounds to 25% across workflows
βlet me take Agent EKIS as an example. It increases complexity in two levels. It increases essentially technical complexity. How do I manage essentially accuracy of a process with multiple agents interacting with each other, each of which is probabilistic and a 1% error may propagate to a 25% error at the end of the process? So how do I organize for this, which is the technical complexity? The second complexity is organizational complexity. We are very excited about agents, but what does the organization structure of the future look like? Like, do we have agents who are acting as employees working together with humans?β
80% of tech transformation challenge is organizational, not technical
βFor every generation of new frontier technologies, as hard and as fun as it is to talk about the speeds and feeds and all the technology, almost 80% of the challenge is on the organizational side. It is around change management. It is around aligning for digital, for AI. We've talked about rewiring organizations, and it is matching your technology strategy to your corporate strategy. It's having the right talent in place. So the tech stack is important, having the data is important. But then how do you change your incentives? How do you change your culture? How do you change the way you operate?β
SpaceX is manufacturing custom GPUs for internal workloads
βSpaceX is telling its investors it wants to start building its own GPUs. This is a massive deal, because I think this is really just showing us how tight the compute market has gotten. You have like a rocket company that is looking at building its own silicone. I think this is what is, what this is telling us right now is that we've basically reached the point where the biggest tech companies in the world no longer trust the GPU supply chain enough to just be customers.β
Liquid cooling is the hidden gem powering AI data center growth
βwhen we think about cloud and edge computing, five or six years ago, we spent a lot of time thinking about it. One thing I'd like to say is those scientists and engineers have not stopped thinking about innovation. For example, as we think about GPU-based clouds, AI-based data centers, there's some really interesting advances in liquid cooling. You might say, well, why is that a hidden gem? Well, when you start to have the number of data center growth that we're anticipating, I believe data centers are growing 30 percent year-on-year. I mean, it's an astounding trend of demand, especially if I just think these AI systems are hungry hippos. They need a lot of compute power. That's really hot.β
Don't wait on the sidelines in this technology era
βDon't wait on the sidelines. A lot of times business leaders say, wow, that's really interesting. Maybe I'll watch someone else run that play first, and then I'll be a fast follow. Quite frankly, there have been many technology eras where that is absolutely and has proven out to be the right thing to do. But in this instance, over the last couple of years, really since Chat GPT was released, I don't think you can wait on the sidelines. So you just got to get in the game and run the ball, and you got to compete to win.β
βJeremy explains why traditional banking fails to support the needs of AI agents, and how stablecoins like USDC facilitate an internet-native economy.β
Google Cloud launches a dedicated enterprise agent platform
βGoogle just had their big cloud event in Las Vegas this week. They rolled out the Gemini Enterprise Agent Platform. I think this is basically their attempt to take a swing at OpenAI and Anthropic in the enterprise for the AI agent race. I think first, access to over 200 models through Model Garden is something that they're going to be rolling out. So enterprises aren't going to be locked into just Gemini.β
SpaceX is manufacturing custom GPUs for internal workloads
βSpaceX is telling its investors it wants to start building its own GPUs. This is a massive deal, because I think this is really just showing us how tight the compute market has gotten. You have like a rocket company that is looking at building its own silicone. I think this is what is, what this is telling us right now is that we've basically reached the point where the biggest tech companies in the world no longer trust the GPU supply chain enough to just be customers.β
βThe model itself, basically OpenAI's positioning on this is that it is quote unquote a fully agentic model, meaning that it's designed to complete all these multi-step computer tasks with minimal human direction. They specifically highlighted five different categories, analyzing data, writing and debugging code, operating software directly, researching online, creating documents and spreadsheets autonomously.β
GPT 5.5 pricing introduces a high-tier pro subscription
βThe pricing, I think, is also where this gets a lot, gets very interesting. The standard GPT-55 is $5 per million input tokens and $30 per million output tokens. That doubles GPT-54 on paper, but OpenAI is claiming that the model uses tokens more efficiently, so the real world cost per task should be roughly flat or better. GPT 5.5 Pro is $30 in and $180 out, which is very expensive.β
Tech giants leverage AI to restructure white-collar workforces
βMeta has just sent a memo to all of their employees, announcing that they're cutting about 10% of their workforce. This represents about 8,000 people. Zuckerberg basically told everyone this at the start of the year, when he said 2026 would be quote, the year AI starts to dramatically change the way we work. I think this is a story about how AI is starting to really reshape payroll at the biggest companies on the planet.β
Agentic AI marks the shift from chat to autonomous action
βdid we find some things that really popped this year that we, for example, added agentic AI? Again, we don't think this is just going to be a one-year thing, but I think all of us have been talking about agents a lot quite recently. And we think we're just starting on that journey and many of our clients are just starting on that journey. So certainly while AI has been with us all the years that we've been studying trends, this year the idea of agents, this idea of foundation models acting in the world, doing things that are more autonomous and more multi-step, that's something that clearly came to the surface here as well.β
Anthropic Mythos leads the new AI cybersecurity race
βMicrosoft is actually integrating that, but it's just a preview. It's directly inside of their secure coding framework, and the idea is that Clawed gets used for threat detection, vulnerability scanning, and incident response inside of Microsoft's DevTooling. And at the same time, OpenAI has been briefing US federal agencies, state governments, and also the Five Eyes intelligence partners on a version of their model called GPT 5.5 Cyber.β
Bioengineering extends far beyond pharma into food and fabric
βif you look at the future of bioengineering, for instance. Now, by the way, that's not just life sciences companies, it's not just pharma and cynical companies. I mean, the amazing thing is when you can program life, you can start to create new materials. And so what does that mean? Everything from apparel, artificial spider silk in order to create fabrics, to agriculture, again, incredibly important all across Asia, for instance, and whether or not it's being able to develop crops that can deal with climate change and or just increase the nutrition and as well as the taste of different crops.β
Agentic banking will accelerate the disruption of incumbents
βFintech is only three to five percent of the market for financial services, and we've been talking for a long time about how banks are going to have to wisen up because startups are coming to eat their lunch. But the rate of that flip now with agents opening bank accounts is making traditional banking look like primitive technology. The next two to three years will be unlike anything we've seen.β
Tech giants leverage AI to restructure white-collar workforces
βMeta has just sent a memo to all of their employees, announcing that they're cutting about 10% of their workforce. This represents about 8,000 people. Zuckerberg basically told everyone this at the start of the year, when he said 2026 would be quote, the year AI starts to dramatically change the way we work. I think this is a story about how AI is starting to really reshape payroll at the biggest companies on the planet.β
Autonomous vehicles show how fast we normalize new tech
βLareina, I live in San Francisco, and this is a little bit about the way the technology trends feel. You sort of have to look back because things become so normalized so easily. So there are a bunch of autonomous vehicles driving around San Francisco, and it used to be like, wow, that car has no driver, right? And I've told this story before, but our young daughter saw a taxi go by, and I had to explain to her what a taxi was. And she said, oh, that's like a Waymo with a driver.β
Google Cloud launches a dedicated enterprise agent platform
βGoogle just had their big cloud event in Las Vegas this week. They rolled out the Gemini Enterprise Agent Platform. I think this is basically their attempt to take a swing at OpenAI and Anthropic in the enterprise for the AI agent race. I think first, access to over 200 models through Model Garden is something that they're going to be rolling out. So enterprises aren't going to be locked into just Gemini.β
Financial infrastructure must remain agnostic across AI models
βWe went heavy with the Claude branding because they seem to be the far and away winner right now, especially for businesses, but we built it to be agnostic behind the scenes. You're exactly right; even though our branding is currently Claude-oriented, it is agnostic. I would be terrified to just be a pure software company right now.β