AJ, ETAPX's founder and CEO, sat down with PX News for a wide-ranging conversation about where artificial intelligence and the broader tech industry actually stand right now — not the keynote-stage version, and not the version from someone convinced the whole thing is a bubble. He talked through what he thinks most of the AI industry still gets wrong about this exact moment, how that read shapes what ETAPX chooses to build next, and previewed the energy behind a run of new product releases headed out from the company over the coming months, several of which will get their own dedicated coverage here on PX News. This is that conversation, mostly in his own words.
The Moment We're Actually In
Ask AJ for his honest read on artificial intelligence right now and you don't get a keynote answer. You get a pause, and then an argument. "Every few months there's a new consensus about where AI is headed, and most of those consensuses are wrong in the same specific way," he says. "They extrapolate from whatever happened last week instead of asking what's actually going to hold up. That's not analysis. That's just recency bias with better production values."
His actual view sits in the uncomfortable middle that doesn't get much airtime. It isn't the breathless version, where AI is about to remake every industry within the year. It isn't the dismissive version either, where none of this is real and everyone talking about it is selling something. What he thinks is happening is narrower, and in his telling, more useful: the raw intelligence of frontier models has gotten good enough, and cheap enough, that having access to a capable model stopped being the advantage a while ago. Almost anyone can call a good model now. What's actually scarce is everything you build around it — the judgment about what to build, the discipline to build it well, and the follow-through to ship it and keep improving it instead of moving on to the next demo.
"The industry spent two years arguing about whose model was smartest. I think that argument is basically over, or at least it's not the argument that matters anymore. The interesting question was never 'which model is smartest.' It's 'what did you actually build with it, and did you build it well enough that someone's life is better for using it.' Almost nobody is asking that second question loudly enough."
— AJ, Founder & CEO, ETAPX
That distinction shapes almost everything ETAPX has chosen to prioritize this year. If raw model access were still the scarce resource, the winning move would be chasing whichever lab published the newest benchmark and wiring it in first. Since it isn't, AJ's attention has moved almost entirely toward the layer above the model — the actual product, the actual workflow, the actual experience of a person trying to get something done. "We use plenty of frontier models. Everyone does at this point, or they will soon enough. That was never going to be the whole company. It's the material, not the building."
Signal, Noise, and the Discipline to Tell Them Apart
Part of what makes this moment disorienting, AJ argues, isn't a shortage of information about AI. It's an overwhelming surplus of it, most of it competing for attention rather than trying to be accurate. New models ship on a weekly cadence. Benchmarks get published, cherry-picked, and re-litigated within days. Every announcement arrives dressed as a breakthrough, whether or not it actually is one. "I don't think people outside this industry understand how much of what gets called news is just marketing wearing a lab coat," he says. "Doesn't make the underlying work fake. It just means you have to get better at telling the two apart, because nobody's going to do that sorting for you."
What Actually Matters When Everything Is Loud
AJ's own filter for cutting through that noise runs close to the instinct behind GLSRM, ETAPX's intelligence platform for the AI industry, though he's careful to draw a line between the product and his personal read. "GLSRM exists because I got tired of not being able to tell the difference between a real capability jump and a good demo video. That problem was real for me before it was a product for anyone else." His own rule of thumb hasn't changed much since: trust standings tracked over time more than any single announcement, trust what a system can actually do in someone's hands over what a launch video implies it can do, and stay deeply suspicious of anything that only makes sense in the seven days after it launched.
He's equally direct that this isn't a call to cynicism. "The failure mode isn't just falling for hype. It's overcorrecting so hard that you dismiss everything, and then you miss the two or three shifts a year that are actually real. I've watched smart people do both. The discipline is staying open enough to notice something real, and skeptical enough not to get played by something that isn't."
Agentic Work Is the Real Shift
Pressed on what he thinks the one actually-real shift is right now, AJ doesn't hesitate. "It's agentic. Not as a buzzword — I mean the specific, boring-sounding thing where a system can hold a task, take real steps toward finishing it, and check back in with you instead of handing you a wall of text and calling it done. That's the part that was missing for years, and it's the part that finally started working."
That shift is the entire reason Influxx exists, and AJ says the timing wasn't a coincidence — it followed directly from watching the underlying capability cross a threshold worth building a whole product around. Influxx is ETAPX's flagship agentic environment, a real place to hand off actual work and have it move forward without being babysat every step of the way, built around the Agentic Dev Environment, or ADE, along with an ADE mobile app for checking in on tasks from a phone, and Podz, the workspaces where that work actually lives and gets organized. "Influxx isn't a chatbot with a task list bolted on. It's built around the idea that you should be able to hand something off, walk away, and come back to real progress. That sounds simple. It took a long time to be true."
Why Influxx Is Where the Energy Is Right Now
ETAPX just shipped native macOS and iOS apps for Influxx — a release covered in its own detail elsewhere on PX News — and AJ points to it as a preview of the pace, not a peak. "We didn't build native apps because it looked good on a roadmap slide. We built them because people were already stretching the web app past what a browser tab should reasonably be asked to do, and the honest answer was to give the environment a real home on the platforms people actually live on. That's the posture going forward, not a one-time move." He's blunt that the ADE, the mobile app, and Podz will all keep getting more central to how people use Influxx, not less. "Agentic work is the shift I actually believe in. We're going to keep building like we believe it."
"If you want to know what I think matters most in AI right now, don't look at what I say in an interview. Look at where we're putting the people. Right now, more of this company is working on agentic product than on anything else we've ever built, by a wide margin. That's not a hedge. That's a bet, and it's the one we're most sure of."
— AJ, Founder & CEO, ETAPX
What's Coming From ETAPX: A Preview of New Product Releases
This is the part AJ was most willing to talk about in general terms and least willing to fully spell out — on purpose. "I'd rather give you the honest shape of what's coming than a watered-down version of the details, before the details are actually locked. Some of this has its own article coming. I don't want to spoil those by rushing the specifics here." What follows is the shape, not the specifics.
- Influxx and the agentic core, deepened further: the native macOS and iOS apps were a platform move, not a finish line — expect the ADE, Podz, and the mobile experience to keep absorbing more of what used to require the web app open in a tab.
- Going deeper on Apple's platforms: AJ says the recent native release was a first step in a longer commitment to building real, first-class citizens of macOS and iOS rather than treating those platforms as an afterthought behind the web product, with more to say about that commitment soon, in its own dedicated piece.
- A new category we're exploring: ETAPX is looking seriously, and honestly, at what a physical device built around agentic work could look like — early enough that AJ is deliberately not attaching specifics to it here, with a separate article to cover that decision properly once there's something real to say.
- The unglamorous work of scaling capacity: more people relying on ETAPX products every week means investing ahead of that curve rather than catching up to it after something breaks, work that rarely makes a headline but that AJ says matters as much as anything on this list.
Asked why he's willing to gesture at things the company isn't ready to fully announce, AJ's answer loops back to a posture he's talked about before. "I'd rather tell you honestly that something's coming and that I'm not ready to give you all of it yet, than stay quiet and let it just appear one day with no context. People can handle 'we're working on it, more soon.' What they can't do anything useful with is silence."
Why We're Willing to Say This Early
Of everything on that list, the new category draws the most questions internally, and AJ knows it'll draw the most outside the building too. He's careful with how far he takes it in this conversation, and equally careful to explain why he's willing to mention it at all before it's fully ready. "I could have just not brought it up until the day we were ready to announce something specific. I thought about that. I decided against it, because it's not how we've done anything else at this company, and I wasn't going to make an exception just because this one felt bigger."
What he will say is philosophical rather than technical — intent, not specification. "We're exploring what a physical device built around agentic work could look like. That's genuinely as far as I'll take it today, not because I'm being coy, but because anything more specific than that isn't real yet. We don't have a spec sheet I'm hiding from you. We have a question we're taking seriously: if agentic work is really the shift I think it is, does it deserve a home that isn't just a laptop screen or a phone you're already using for forty other things?" A separate, dedicated article will cover that exploration properly once there's something concrete to cover — this is context for how it fits into AJ's broader read on the moment, not the announcement itself.
"I'd rather say 'we're exploring this, honestly, and we're not done thinking about it' than either stay silent or overpromise something we haven't earned yet. Both of those are more comfortable than what I just told you. I don't think comfortable is the same thing as honest, and I'd rather be the second one."
— AJ, Founder & CEO, ETAPX
He's equally clear about what this isn't. It isn't a pivot away from software, and it isn't a signal that Whistlr, GLSRM, Ocsidian, or Influxx are taking a back seat to anything. "Nothing about looking at a new category means we're less committed to the products people already rely on. It means we think the agentic shift might be big enough to eventually need more than a screen. That's a hypothesis right now, not a roadmap. I'll tell you when it's a roadmap."
Scaling Without Losing the Thing That Makes This Work
The least exciting item on AJ's list is, by his own admission, the one he thinks about the most: making sure ETAPX can actually support what it's building before it ships more of it. "Every one of these things — more Influxx, going deeper on Apple's platforms, a new category, all of it — is worthless if the ground underneath it can't hold the weight. I've watched other companies ship something great and then let it fall over the moment real numbers of people showed up. I think about that more than I think about almost anything else."
That translates into decisions that rarely show up in an announcement: hiring ahead of need instead of behind it, investing in capacity before it's an emergency, and saying no to shipping something faster if shipping it faster means shipping it thin. "Craft and capacity aren't in tension the way people assume. You can't actually have craft at any real scale without capacity behind it. A beautifully designed thing that falls over under real usage wasn't actually well-crafted. It was well-crafted for a demo."
The Trade AJ Keeps Making on Purpose
Pressed on whether that caution is in tension with how much he previewed earlier in this same conversation, AJ pushes back on the framing. "Telling you honestly what we're exploring isn't the same as rushing it out the door. I can be excited about where this is headed and still refuse to skip the boring work that makes it real. Those aren't opposites. They're the same discipline pointed at two different jobs."
"I'd rather move one step slower and have it actually hold, than move fast and find out in public that it didn't. We've made that trade on purpose more than once this year, and I'll keep making it, even when it's the less exciting answer to give in an interview like this one."
— AJ, Founder & CEO, ETAPX
Frequently Asked Questions
What does AJ think is the biggest misconception about AI right now?
That the interesting question is still "which model is smartest." AJ's view is that frontier AI has gotten capable and cheap enough that access to a good model stopped being the scarce resource. What's scarce now is what a company actually builds around that capability, and how disciplined it is about shipping that well instead of chasing the next headline.
What new product releases is ETAPX working on right now?
AJ previewed several directions without fully detailing any of them: continued investment in Influxx's agentic core, including the ADE and Podz; a deeper commitment to Apple's platforms following the recent native macOS and iOS apps; early exploration of a new hardware category built around agentic work; and ongoing investment in the infrastructure and capacity needed to support all of it. Several of these will get their own dedicated articles on PX News as they firm up.
Is ETAPX actually building a hardware device?
ETAPX is exploring what a physical device built around agentic work could look like. AJ was direct that this is early, philosophical, and intentionally light on specifics right now — there's no spec, timeline, or pricing to share, and a separate article will cover the direction properly once there's something concrete to say.
Why is ETAPX putting more investment into Apple's platforms?
ETAPX recently shipped native macOS and iOS apps for Influxx, and AJ describes that as a first step rather than a one-time project. His reasoning is that agentic work deserves a real home on the platforms people already live on, not a browser tab standing in for one, and he's said there's more to share on that commitment soon.
How does AJ think about balancing speed with reliability as ETAPX grows?
He treats it as a discipline rather than a tradeoff, saying he would rather move slightly slower and have something actually hold up under real usage than ship faster and find out in public that it didn't. That means investing in capacity and infrastructure ahead of need, which he considers just as important as any feature on ETAPX's roadmap, even though it rarely makes for an exciting announcement.
None of this was AJ fully unveiling anything. It was closer to the connective tissue between a handful of announcements still on their way from ETAPX — his honest read on where AI and the wider tech industry actually stand, and why that read is exactly what's pointing the company toward Influxx's next chapter, deeper Apple-platform investment, a new category it's willing to say out loud that it's exploring, and the quieter work of making sure all of it holds. The specifics, as promised, are coming. This was just AJ telling you, plainly, why.
