Marc Andreessen: The real AI boom hasn’t even started yet

AI has now proven it can do real reasoning and problem-solving in domains that matter, including developing new math theorems, and the world's best programmers now say AI codes better than they can.

“the answer to that is yes right um and you know the last 12 months and especially the last even just the last three months have really proven that like AI can really do like you know you're seeing it all now you know you can actually you know AI is now developing new math theorems um you know there you know over the holiday break you know there's sort of the what it feels like the AI coding thing you know really hit critical mass uh and the world's best you the world's best programmers right including like Lisbald's you know for the first time over the holiday break basically said yeah AI is now coding better than we can”

AI models are currently testing at IQ levels around 131-140, approaching the 160+ level associated with paradigm-shifting scientists like Einstein and Feynman, with no theoretical upper bound on AI intelligence if biological constraints are removed.

“existing AI models right now are kind of testing around the 131 140 level... they're arguably on the mass high starting to get to the 160 level now. But like I I think we're going to have AI models relatively quickly that are going to be like 160, 180, 200, you know, 250, 300... And I think that's great, right? Like I feel I feel as great about that as I do about the fact that we occasionally get an Einstein”

The world's best coders are now spending their day 'arguing with AI bots' about code quality, debugging, and specifications rather than typing code by hand, suggesting a shift from code generation to code orchestration.

“if you talk to the world's best programmers today what they'll tell you is oh my job is I'm sitting there and I'm orchestrating 10 code bots right coding bots that are running in parallel right and and literally they sit there and they shift from browser you know browser to browser or terminal to terminal and they're and they're they're watch their their day their day job now is kind of arguing with the AI bots trying to get them to like write the right code”

Many smart AI researchers off the record believe 'there really aren't any secrets among the big labs'—they all have access to the same information and knowledge, and they leapfrog each other regularly.

“like many of the smartest people I know in the field when I when I really kind of talk to them kind of, you know, get a couple drinks into them, they're like, 'Yeah, they're basically, you know, one theory is like there really aren't any secrets among the big labs.' like the big labs kind of all have the same information and they kind of have all the same knowledge and they you know they're kind of they lap each other on a regular basis but you know there's not a lot of proprietary anything at this point”

Current AI models score in the 131–140 range on human IQ tests, are starting to approach 160 (Einstein/Feynman level), and will likely exceed human cognitive ceiling in the near term, which is desirable because more Einstein-level intelligence is good regardless of source.

“existing AI models right now are kind of testing around the 131 140 level... I think we're going to have AI models relatively quickly that are going to be like 160, 180, 200, you know, 250, 300... the world would be better off with more Einstein. And of course the world would be better off with machines that have IQ, you know, more IQ like Einstein are greater than Einstein.”

Wealth inequality is not a risk if AI-driven deflation occurs, because falling prices increase real wealth across society; the issue is whether some people remain unemployed, not whether they are poor.

“in this kind of utopian dystopian scenario that people have, it's not there there's no scenario in which like everybody's just poor. In fact, it's it's quite the opposite, which is everybody gets a lot richer because prices collapse”

The trust in legacy institutions is in full-scale collapse; freedom of speech, thought, and open discussion have been liberated in a one-way direction; and massive geopolitical shifts are happening simultaneously across the US, Europe, China, and Latin America.

“the trust that a lot of people have had in kind of what you describe as kind of legacy institutions around the world is I I think in kind of full scale collapse right now. By the way, there's a lot of data data to support that. And so I think there's just there's there's like a lot of structures and orders and uh institutions that people have just relied on for a long time that have just proven to not be up for the up for the challenge. And then kind of corresponding with that is the national and global conversation have become like let's say liberated.”

The trust that people have had in legacy institutions around the world is in full-scale collapse right now, there are real structural impediments in the economy and political system that prevent rates of change anywhere near the rates we had in the past, and this is comparable in magnitude to the fall of the Berlin Wall in 1989 or the end of World War II.

“the trust that a lot of people have had in kind of what you describe as kind of legacy institutions around the world is I I think in kind of full scale collapse right now... these like incredibly massive geopolitical shifts that are happening... I think a lot of assumptions are being pulled out in the into the daylight and and re-examined... those three kind of big mega things are kind of all colliding um at the same time... comparable in magnitude to maybe the fall of the Berlin Wall in 1989... maybe the end of World War II”

AI is the philosopher stone, a technology that transfers sand—the most common thing in the world—into thought, the most rare thing in the world.

“AI is the philosopher stone. Now we have a technology that transfers the most common thing in the world which is sand converted into the most rare thing in the world which is thought.”

The real-world cost of housing, healthcare, and education will collapse once AI reduces production costs, effectively giving everyone a raise and increasing spending power to support new economic sectors.

“by the way, if you to the extent that you do have unemployment coming out the other side of that, it's it's now much cheaper to provide the kind of social safety net to prevent people from being emirated, right? Because the prices of all the goods and services that like a welfare program has to pay from, they're all collapsing, right? And so the price of healthcare collapses, the price of housing collapses, the price of education collapses, the price of everything else collapses because this this this this incredible impact that AI is having.”

Despite the speed of AI model commoditization (GPT-3 clones emerged within a year, open-source versions on fraction of hardware, DeepSeek replicated US lab advances rapidly), it is not yet clear whether moats exist in AI models themselves, and defensibility could accrue at the application layer instead if domain-specific adaptation matters more than base model capability.

“even at the level of like LLMs or you know AI models like you can squint and make that argument either way... it's like okay we need to harness the base model as kind of the engine into a into a domain involving human beings u where you need to like actually have it fit for purpose... maybe the LLM's commoditizing maybe the value goes to the apps um and and and again you can kind of squint either way on that one”

Claude Code building co-worker in a week and a half is both impressive (showing Claude's capability) and concerning (suggesting low barriers to entry if you can build a valuable product in a week).

“co-work was developed in a week and a half [laughter] like like h how much complexity could there be? How much of a barrier to entry can there be in something that was developed in a week and a half?”

Watching how AI agents work and think as they solve problems is an underutilized learning method; observing AI reasoning helps you understand architecture, algorithms, and decision-making in ways that reading code alone cannot.

“Two tricks I've heard along those lines. One is uh to watch the output. What the agent is doing and thinking as it's doing the work. So, if you're not an engineer, is just sit there and watch it think and make decisions. And it's almost become this like layer on top of learning to code is learning to see what the agent is doing and thinking because that teaches you about architecture.”

Voice input and voice AI (Meta glasses, Whisper Flow, voice transcription with LLM understanding) represents a coming wearable revolution that will make AI more natural and ubiquitous than screen-based interfaces.

“I am just completely in love with all the AI voice stuff. Um I think it's just absolutely amazing... the all the wearables, like all that stuff is going to be big. The meta glasses um, I think there's going to be a whole wearables revolution here.”

AI is the 'philosopher stone' because it converts sand (the most common thing in the world) into thought (the most rare thing in the world), fulfilling the alchemical goal that eluded Newton and early scientists for centuries.

“AI is the philosopher stone. Now we have a technology that transfers the most common thing in the world which is sand converted into the most rare thing in the world which is thought.”

Silicon Valley has been characterized by nine major technology platform waves (Silicon manufacturing, personal computers, internet, mobile, cloud, and now AI, among others), and the region's strength is 'indeterminate optimism'—betting on many founders trying many approaches rather than predicting winners.

“AI is the ninth major technology platform in the history of Silicon Valley... Silicon Valley is still called Silicon Valley... now we're on like wave nine... the company town phenomenon where the company is the industry like the indeterminate optimism... the ecosystem flexibility of the ecosystem met that the the Silicon Valley could could morph um into all these categories”

What happens when you're good at two things (like design and coding) is the additive effect is more than double, and when you're good at three things the additive effect is more than triple, making you a super relevant specialist in the combination of the domains, as Scott Adams demonstrated by combining cartooning and business understanding to create Dilbert.

“the additive effect of being good at two things is more than double. The additive effect of being good at three things is more than triple. You become a super relevant specialist in the combination of the domains... Scott Adams he used to say he said um you know I I could have been a pretty good cartoonist um or I could have been like pretty good at business but the fact that I was a cartoonist who understood business made me like spectacularly great at making Dilbert, right?”

Confident predictions about AI industry structure (which company will dominate, where moats will exist, what the killer app is) in real-time suffer from the same systematic failures as predictions about the internet in 1993–2010; the right strategy is to place many bets and remain adaptable rather than to forecast the winner.

“if you look back on those predictions a few years later and you you can do this by the way if you pull up like coverage of the internet from like 1993 through like 1997 or even through like for that matter even through like 2005 or 2010 and you look at like the kinds of confidence statements people were making in the first 10 or 15 years like I would say like almost all of them were wrong”

Using multiple AI models in competition or debate (LLM councils) to critique each other's work can surface errors and insights that single-model outputs miss.

“you can have one AI write the code, you have another AI debug the code, and so you can actually use you can play the AIs off against each other and get them to argue with each other. Um, and yeah, the these are all these are all the kinds of skills that are going to become, I think, incredibly valuable.”

The most leading-edge AI founders are thinking about whether entire companies can exist where the founder does everything by overseeing an army of AI bots, potentially creating one-person billion-dollar companies.

“can you have entire companies where you have basically the founder does everything right because what the founder is doing is like overseeing an army of AI bots and and there's sort of this you know there's kind of this holy grail in our industry that's been running for a long time which is like can have the can you have like the one person billion dollar outcome”

There is a 'Mexican standoff' between product managers, engineers, and designers where each now believes AI enables them to do the other roles, making all three roles simultaneously seem redundant and creating an opportunity for superpowered individuals who combine all three skills.

“There's like a Mexican standoff happening between those three roles. Every coder now believes they can also be a product manager and a designer because they have AI. Every product manager thinks they can be a coder and a designer. And then every designer knows they can be a product manager and a coder. They're actually all kind of correct.”

Leading edge AI founders are exploring three layers of AI impact: redefining products themselves (Nano Banana vs. Photoshop), empowering jobs (AI-enabled workers), and reconstituting what a company is (one-person companies, fully AI companies, autonomous AI agents on blockchain).

“I think there's like three layers of it... layer one is they're thinking all right how how does AI redefine the products themselves... the next layer is actually a lot of what we've already talked about which is AI changing the jobs... and then I think the third shoe to drop hasn't quite dropped yet, but it's it's you know it's kind of the big one which is like all right like the the the basic idea of having a company right you know does that change”

Direct exposure to domain practitioners sharing knowledge through podcasts, newsletters, and Substack is dramatically underrated compared to traditional mediated media (TV, newspapers, magazines), and is a primary source of alpha in understanding emerging fields.

“the actual practitioners in the field who are actually creating content, I think probably is still like dramatically under underrated... the reason for that is like we we're we're used to being in this mass media kind of culture in which basically everything is mediated, right?... now more and more it's just no you actually want like smart people who are actually working on something explaining themselves”

The conversation between young hip progressive (Pedro Pascal) and crusty right-wing sheriff (Joaquin Phoenix) in the film Edington, played out against COVID, BLM protests, and tech anxiety, captures what it means to be human in the 2020s by showing how people experience real-world events primarily through internet and social media.

“the reason I love the movie so much is... it's the first movie that does a really good job of showing what it what it was like especially in that era to live in a world in which there were things happen in the real world and people were kind of experiencing events online, you know, like in a way that was like very central in their lives, right?”

Great designers (capital-D design) focus on high-level questions (purpose, human fit, emotional resonance, life integration) while AI handles task-level design (icon generation, UI assembly), elevating designers from execution to strategy.

“the the the task level of like design the perfect icon, right, is going to be like all right, the A&'s going to do that all day long... But like what are we trying to do? Like the, you know, kind of capital D design of like, all right, what is this thing for? And how does this how is this going to function in a world of human beings?”

People who build with AI by leveraging it to expand laterally into adjacent domains become non-fungible; their scarcity and unique skill combinations make them immune to replacement, unlike specialists in single domains.

“the key for career planning is he said don't be funible, right? And you know that's he's an economist and so that was economics speak and and what that means is what that means essentially is don't be replaceable. And so don't be a cog. Right? So, and what that meant was don't just be one thing... if you have this if you have this combination of things that's actually quite rare, then all of a sudden you're not fungeible. Not not only you're not funible, like you're actually massively important”

When AI-generated work fails or doesn't meet expectations, asking 'What could I have said differently to avoid this error?' teaches more than receiving the wrong result, because it illuminates the gap between your intent and your expression.

“The other is uh a couple podcast guests have mentioned this. When you get stuck and then you figure out how to unstuck yourself, you ask it, 'What could I have done differently? What could I have said that would have avoided this error in the first place?'”

Silicon Valley operates as an ecosystem/company town where the company is the industry itself, enabling continuous reinvention across nine major technology platforms without central planning, demonstrating the power of indeterminate optimism and ecosystem flexibility.

“AI is the ninth major technology platform in the history of Silicon Valley... Silicon Valley is still called Silicon Valley. We haven't made Silicon here in decades... the the the company town phenomenon where the company is the industry... the the again the indeterminate optimism the nobody had nobody had to sit and plan and say okay in the 1990s Silicon Valley is going to do the internet in the 2000s they're going to do the smartphone in the 2010s they're going to do the cloud in the 2020s they're going to do AI”

If AI causes massive productivity growth, the necessary economic consequence is price deflation across AI-affected sectors, which is equivalent to a giant raise for everyone, because goods that cost $100 now cost $10 or $1, expanding purchasing power and enabling social safety nets to be more affordable.

“The necessary economic calculation of what happens is massive massive productivity growth. The consequence of massive productivity growth, what that literally means mechanically is more output requiring less input, right? So you get more economic output for less input... you get lots of goods and services in all those affected sectors. The result of those gluts is you get collapsing prices, right? The collapsing prices mean that the thing today that cost you $100 now cost you $10 and now cost you $1. That's the equivalent of giving everybody a giant raise, right? Because now they have all this additional spending power.”

The term 'fungible' (replaceable/interchangeable) versus 'non-fungible' is a useful mental model for career planning: being a single-skilled cog makes you replaceable, while being a rare combination of skills makes you invaluable.

“my my friend Larry Summers had a had a different version of the Scott Adams thing which is he he used to tell people he said the key for career planning is he said don't be funible, right? And you know that's he's an economist and so that was economics speak and and what that means is what that means essentially is don't be replaceable. And so don't be a cog. Right? So, and what that meant was don't just be one thing, right?”

If AI triples productivity growth in the economy, it would take us back to the same level of job turnover that was happening between 1870 and 1930, when people perceived the world as full of opportunity and new careers, not as an apocalyptic job-loss scenario.

“even if AI like triples productivity growth in the economy, which would like be a massively big deal, it would take us back to the same level of job turnurn that was happening between 1870 and 1930. And if you go back and you read accounts of 1870 to 1930, people just thought the world was a wash with opportunity.”

The history of programming shows repeated cycles of abstraction (calculator → machine code → assembly → C → scripting languages → AI coding) where each layer of abstraction initially faces skepticism but ultimately expands the field and shifts labor to higher-order tasks, and AI coding is the next such layer.

“we've gone from a world in which you literally have people doing mathematical equations by hands by hand. Then we got the first computers... Then we got actually this big breakthrough which was called assembly language... when I was coming up it was higher level languages like C... when scripting you know when scripting languages you know we developed JavaScript at Netscape... there was this big fight in in the technical community which is is scripting real programming or not... of course the answer is yes it very much counted and now most coding is done with the scripting languages... AI coding is the next layer on that”

Design—the capital-D decisions about what something is for, how it fits into human life, what makes people happy using it—will become more valuable as AI automates the task-level work (icon design, visual polish), concentrating human designer effort on higher-order questions.

“the task level of like design the perfect icon, right, is going to be like all right, the A's going to do that all day long... But like what are we trying to do? Like the, you know, kind of capital D design of like, all right, what is this thing for? And how does this how is this going to function in a world of human beings?... the job of designer right will involve much more of those higher level more important components”

Agency—the ability to take initiative, take responsibility, and be a primary participant in events rather than just following rules—has diminished in culture over the last 30 years but is critical for kids to develop in the AI era, as AI is the ultimate lever for someone with agency to fully participate and contribute.

“there's something to be had... in order to lead, you must first learn to obey... but yeah no look there there is like a huge b there's just a huge premium in life on being somebody who is able to like fully take responsibility for things fully take charge run an organization lead a project create something new um and you know maybe yeah that that has been maybe a little bit diminished in our culture over the last 30 years”

In predicting technology outcomes, people state conclusions with excessive confidence that prove spectacularly wrong 3-5 years later; history of internet predictions (1993-2010) shows nearly all confident forecasts were wrong, suggesting humility is warranted on AI moats and outcomes.

“if you look back on those predictions a few years later... the kinds of confidence statements people were making in the first 10 or 15 years like I would say like almost all of them were wrong... again generally like quite badly wrong”

Peter Thiel was more right than I was in our past debate about technological progress—he correctly identified that progress in 'atoms' (physical infrastructure, real-world construction) has been absent for 50 years, while I was focused on progress in 'bits' and missed the broader point.

“I have come much more around to Peter's point of view... the real form of what Peter was arguing was we have lots of process in bit. We have lots of progress in bits, right? But we have we have very little progress in atoms, right?... I think I I was a little bit I don't know missing that or kind of you know kind of glossing that over a little bit”

There is a 'Mexican standoff' between product managers, engineers, and designers where each group now believes they can do the other two roles using AI, but they are all actually correct that AI enables multi-role capability.

“there's like a Mexican standoff happening between those three roles... Every coder now believes they can also be a product manager and a designer because they have AI. Every product manager thinks they can be a coder and a designer. And then every designer knows they can be a product manager and a coder. They're actually all kind of correct.”

The best way to improve one's skills in an AI era is to spend every spare hour in conversation with AI systems, asking them to teach you new domains, critique your work, and assign you problems—leveraging AI as both worker and tutor simultaneously.

“people who really want to improve themselves and like develop their career should be spending every every spare hour in my view at this point talking to an AI being like, 'All right, train train me up like tell me tell supermpower me, tell me how to, you know, train me train me how to be...' It will happily do that.”

AI will make people who are already good at something very good at it, creating a superempowered individual effect, while also creating a divergence where really great people become spectacularly great, not just incrementally better.

“AI is going to take people who are good at doing things and it's going to make them very good at doing things right and so It's going to be a tool that's going to sort of raise the average kind of across the board... there's this other thing that's happening which we're also starting to see and we're really seeing it particularly in coding right now. Um where the really great people are becoming like spectacularly great, right?”

A one-person billion-dollar company is theoretically possible with AI (Bitcoin/Satoshi is precedent), but practically faces friction from edge cases, support, bug fixes, and regulatory/operational tasks that are harder to automate than core product development.

“Bitcoin's Satoshi pulled it off... the open source community, you know, like does that count? I don't know. I guess guess it counts... there's so many little annoying things that I have to deal with with just support tickets and issues and bugs and like it's hard for me to imagine actually a oneperson billion-dollar company”

Agency—the ability and willingness to take initiative, participate as a primary actor in events, and break rules when appropriate—is increasingly critical for kids but has been de-emphasized in schools focused on rule-following, making its cultivation essential in an AI era.

“the thing that you for example you want to train kids to do is like follow all the rules... the school system, the K through2 school system or whatever has gotten kind of more and more focused on that over time. And it's like yeah, it's like no, you you should actually... at unit unit n equals one, like of your kid... there's there's something to be had... You know, you need to keep keep him with some level of structure in his life and not just and not just pure agency but yeah no look there there is like a huge b there's just a huge premium in life on being somebody who is able to like fully take responsibility for things fully take charge run an organization lead a project create something new”

Two experiences characterize intellectual limitation: (1) having insufficient time, energy, or memory to complete cognitive projects; (2) encountering people who are demonstrably smarter and who will consistently outthink you, creating a sense of intellectual inadequacy.

“I have this experience all the time. Well, two two experiences... one is just like like I'm just like like I know I ought to be able to do this, but like I just can't... the other is... I know a bunch of people who I know for sure are smarter than I am... at a certain point, you know, it's like for the first half of the conversation, I'm just taking notes the entire time. And for the second half of the conversation, I'm just like, Like, me. like this person is just smarter than I am”

People should not read newspapers and magazines from last week or month because almost none of the predictions materialize and reported 'urgent' issues prove irrelevant, making current media a waste of attention except for high-stakes breaking news.

“if you go back and you read old newspapers, and by the way, you can you can do this. Just read last week's newspaper, right?... and just go back and read it and be like, 'Oh my god, like none of this happened.' like n that none of what they predicted played out the way that they said that it would.”

Individual founders need to be determinate optimists with specific plans for what they're building, but VCs can and should be indeterminate optimists because one core virtue of capitalism is having many bright people making different bets.

“the founders need to be deter determined optimist. Like they need to have a very specific plan now. And look, the the critique the critique always, you know, the critique from the founders is, oh, UVC's have it easy because like you don't have to like you don't actually have to commit, right? You don't actually have to like make you you don't actually have to like, you know, you don't have to make the bed you lay in. You can like place multiple bats. you can operate a portfolio, you know, you should have a lot more sympathy for us as founders, you know, because we, you know, we only get to make the one bet. Um, you know, and there's there's truth to that. You know, the counter-argument on that is the founders get to run their companies. We don't.”

The additive effect of being good at two things is more than double, and being good at three things yields more than triple value, creating a 'super relevant specialist' in the combination of domains—Scott Adams demonstrated this with Dilbert by being both a good cartoonist and understanding business.

“the additive effect of being good at two things is more than double. The additive effect of being good at three things is more than triple. You become a super relevant specialist in the combination of the domains.”

To truly leverage AI as a coder, you must still understand the full technical stack from assembly and machine code to network architecture, because you need to evaluate and debug what AI generates, and this depth is what enables superpowered coders to be 10x or 100x more productive than mediocre AI-assisted coders.

“if you want to be one of the best software people in the world and I want to build new software products and technologies that like really matter then yeah you 100% want to still be you want to go all the way down you want your skill set to go all the way down to the assembly to assembly and machine code you want to understand every layer of the stack”

The primary frustration of intellectual work is cognitive and memory limitations (can't do the math in head, can't retain information from 10 books, can't remember learning, encounter people smarter than you), and AI machines without these limitations will provide tremendous relief and expanded capability.

“I just like like I I don't have the eight hours or or by the way the eight weeks or the eight years, right? And like I just don't know enough yet and I'm just like I can't do the math in my head and my memory isn't perfect and like I can't remember and I read you know after you had this you get interested in something you read 10 books and then you're like I forgot almost everything that I just read.”

A barbell media diet (reading current events XOR 50-year-old books, avoiding mid-cycle publications like magazines) avoids recency bias and filler, because old newspapers and magazines are filled with predictions that never materialized, proving the value of focusing on timeless content or real-time signals.

“I have like a almost a perfect barbell strategy um which is I read X and I read old books right so it's basically either like up to the minute what's happening right now um or it's like a book that was written 50 years ago that has stood the test of time... everything in the middle I'm always like much more skeptical about”

One should be very flexible and adaptable at present rather than trying to make confident structural predictions about the AI industry, because massive technological transformation involves complex adaptive systems with many unknowns.

“I think we just need to I don't know my views I my view I need to put like a big discount on my forecasting ability on this one... for me it's much less interesting to try to say okay as a consequence industry structure in five years is going to be X... I think a much much better use of my time is is being being very flexible and adaptable at a time like this”

The concept of AGI as 'human-equivalent capability' will be a footnote moment (e.g., Tuesday in 2026) because the relevant question shifts immediately to what becomes possible when you have machines better than human in intellectual domains, not when you first reach human parity.

“I think this idea of like human equivalent is just going to be like a footnote. It's like, oh yeah, that was just on Tuesday, you know, in in 2026 is when they hit that and it kind of didn't matter because the the next question was like, okay, what are we gonna what are we gonna what do we get to do in a world in which we actually have machines that are better than that”

Homeschooling his 10-year-old combined with heavy use of AI tutoring is Andreessen's chosen method for education because it allows personalization, agency, and leverage of the philosopher stone technology in ways traditional school systems cannot.

“we have a 10-year-old and so, you know, we and we actually homeschool and so we we think a lot about this... a primary thing that we want to make sure to to do is to make sure that he knows fully how to leverage and and get and get benefit out of the philosopher stone, right? Which is uh you know which is to say AI”

Young people today should spend every spare hour talking to AI, asking it to train them and teach them new skills across domains, because AI's teaching capability is underutilized and represents a latent superpower for skill acquisition.

“people who really want to improve themselves and develop their careers should be spending every spare hour in my view at this point talking to AI being like, 'All right, train me up.' Today my guest is Mark Andre... People aren't fully grasping how much this changing. And people who really want to improve themselves and develop their careers should be spending every spare hour in my view at this point talking to AI being like, 'All right, train me up.'”

Concern that 'people aren't grasping how much this is changing' is specifically focused on software engineering, where it's 'pretty clear we're going to be in a world soon where engineers are not actually writing code,' which would have been unthinkable a year ago.

“I think people aren't fully grasping just specifically software engineering and how much that is changing. Like it's pretty clear we're going to be in a world soon where engineers are not actually writing code, which I think a year ago we would not have thought.”

Replet (with vibe coding features) is a compelling product for young developers because it combines code execution with AI assistance and creative direction, enabling a child to build Star Trek simulations with LCARS interface design without traditional technical barriers.

“my 10-year-old. Um I my 10-year-old right now is 100% obsessed with Replet. Um and by the way, it was not from me... he he he through no inter interference on my part uh discovered Replet about uh about three months ago and discovered vibe coding and is like completely obsessed”