Does AI Need an iPhone Moment
Is AI catching on with regular people beyond kicking the tires on ChatGPT?
ON THE POD: Episode 120
This week, on the heels of OpenAI releasing Deep Research, we assess whether AI has caught on with regular people beyond kicking the tires on ChatGPT. Maybe the AI Super Bowl ads will make it more appealing. Plus: the memeification and financialization of everything, volatility as the norm, X shows signs of a financial turnaround and alarm clocks with apps and a subscription program.
Watch PvA on YouTube. Or listen on Apple | Spotify | other podcast platforms.
Out Friday AM.
PVA CONVERSATION
Use case or lose case
TROY: Let’s talk about how your routines have been impacted by AI. What has changed materially for you? Are you more productive? Is your life better? Broadly, how does AI disruption compare to the last big technology disruption, the internet?
AI vs Internet
The internet connected us, democratized information access and in so doing transformed personal communications, media and shopping and service delivery and on and on. Reflecting, I see the last 20 years as a series of innovations that built on each other, all transforming a simple communication protocol, HTTP, into a long string of discrete products that pushed increasing sophistication and control into the hands of the consumer — web pages and email evolved from crunchy static pages to slick, interactive, algorithmic social networks, simple ecommerce evolved into global digital marketplaces, creation moved from experts to individuals, from big machines to the phone. Novel decentralized organizational structures and service offerings emerged… Britannica became Wikipedia, taxis became Uber.
Product brands embodying discrete new value propositions defined this 20 year evolution; AOL, Yahoo, Google, Hotmail, Napster, Paypal, Amazon, Myspace, iPhone, Gmail, Ebay, YouTube, Tinder, Facebook, Instagram, Twitter, Snap, Uber, Netflix, TikTok. The internet gave us new products that did new things and enabled us to say, hey this internet is cool and it does new things I like and I can’t live without my iPhone. Swipe right. DM me.
Tool vs product
AI is different. Naturally we bucket these things into the company and product brands that package them, but today AI is a set of nifty technological innovations that make you smarter and more productive should you decide to make it part of your routine. You don’t really need it. AI is a product, but it's a product you have to operationalize to do other things better and to make your world infinitely personalized.
In the long run, the continuum of things it might eventually do is as wide as your imagination; on the far end, AI is an intelligent human replacement and nothing before compares to its disruptive potential. Closer to home, this week AI analyzed and categorized my 2024 expenses off my Amex masterfile, a nice time saver. I used it constantly for lite daily work support; summaries, call recording, 10K analysis, a thesaurus on steroids. BTW, it totally sucked at creating a morning news summary and I wanted it to be good at this. Humans are much better curators. Robots don’t have taste yet.
AI fetches information, personalizes stuff, summarizes and automates basic tasks. AI changes interface. Now we can talk to computers. Deep reasoning capabilities are expanding the use case to more sophisticated retrieval and analysis. OpenAI’s new Operator show’s the beginning of what agents will do when they replace human browsing, like automating a lunch order or booking a flight. Replacing or augmenting things you used to do yourself.
X is a hotbed of AI boosterism and speculative use cases, like it was for crypto. Paul Graham points out that Replit is almost ready to completely replace developers. Someone else got a second opinion from Grok on her daughter's broken wrist. Good thing she did. This guy had the new OpenAI model author a nice PHD level report of the McKinley Tariffs. HubSpot founder, @dharmesh had it write a 11,000 word report on CRM. This week the new Bytedance video model, Omnihuman showed the Chinese can also make amazing fake videos that are practically indistinguishable from the real thing. There’s endless chatter about the death of consultants. Several of my friends are consultants. I hope they get good at operationalizing AI.
Consulting and other professional disciplines like law and banking are a good place to see the seeping disruption, places where knowledge meets reasoning and where human / AI augmentation fundamentally changes the service economics. The Goldman CEO thinks 95% of IPO paperwork can be done with AI. We are beginning to see how AI reshapes consultancies with Palantir, a next generation consulting company that operationalizes a suite of AI powered analytical tools for defense and national security clientele. Next gen defense contractor, Anduril, does the same. Both are on fire.
Brian, you wondered if AI is really changing our day to day. I think it's a good question. This period of change is not about a product that will magically replace Uber or Instagram or your iPhone. But it will change you and your company if you figure out how to operationalize it. Right now it's an input to a superhuman you, not the new thing in your hand.
If there was an iPhone moment, it was the day ChatGPT launched. Note Google search is still growing. Slow overlapping waves of innovation will culminate in the next crest. Will normal people actually care? Eventually, but marketers will have to do a much better job to convince average folks that all of this is worth the time. Maybe the SuperBowl will help.
Brian is skeptical
BRIAN: My recent focus group of Midwesterners and Canadians led me to conclude that while the spaceship might be over the White House, most normal people have not had any kind of epiphany with AI. Perhaps consumer adoption doesn’t matter, or it happens slowly and then suddenly. I heard the crypto people saying that all the time.
I think a lot of what Alex perceives as journalist bias against anything new is simply the experience of falling for the free breakfast timeshare pitch. Yes, this time it does appear the spaceship is over the White House, and yet anyone paying attention can understand the hype industrial complex is working overdrive. Everyone has too much riding on this to be The Next Big Thing: Big tech, the VCs, Wall Street, the media, all of them. So I think it’s natural to wonder, in honor of the Super Bowl, where’s the beef?
I desperately want these AI tools to be integrated into my life and take care of the “cognitive manual labor” that eats up so much of my day. I’m not wowed by ChatGPT helping you sort your AmEx receipts. It’s useful for research, but I don’t think it replaces search or simply the need to read a lot in order to connect the dots. It is a terrible writer. If I give a note that “this reads like ChatGPT,” that is not meant as a compliment. I’m not as thrilled by the notion that I can now talk to a computer and it talks back. That parlor trick wears off.
When is it going to actually book that trip for me? I keep hearing about it. My guess is this is much farther off than the AI marketers are telling us. Bring up the errors, you get told it's your fault for not being a good robot conversationalist. It’s ironic OpenAI called its first agent Operator because that’s what many people repeatedly say when trapped speaking to a robot at a call center.
At most of the private dinners I do with media executives, I ask them to name AI tools they’re using in their work. Beyond ChatGPT and Claude, not much. And maybe this is inevitable. The enterprise has poured a lot of money into this, mostly in the promise of whacking people, and we get to hear about code being written by the bots and the usual use cases. There’s just a huge gap between the hyperventilation around AI and its impact on people’s lived experiences.
Alex sees the inevitable
ALEX: Behavioral adoption always lags behind technical capabilities. Early email users recognized that fax machines were becoming obsolete. Cord-cutters saw that cable boxes were becoming relics. These outdated technologies persist for years, and their momentum can mask the inevitability of change. The process resembles a dam - pressure builds while everything appears stable until rapid change occurs suddenly.
Major tech companies like Google and Apple have marketed the shift poorly. They overuse the term "AI," which research shows doesn't resonate with people, and often showcase unrealistic or off-putting use cases. Their persistent promotion of scenarios like "book a trip" or "write a heartfelt note" either fall short in practice or feel inauthentic. The constant noise about AI as either humanity's savior or destroyer further muddles public understanding of its practical applications. Altman and his groupies spend more time talking about what’s next than what’s now.
Much of this comes from operating in what is very likely an enormous bubble. Lots of potential but also a lot of noise and grandstanding to stand out from the crowd. Discussions about AGI may drive investor FOMO but don't help consumers understand the technology that’s out today. To increase adoption, we should focus on utility and developing new user behaviors. Today's tools are remarkable and transformative. If we stepped back to evaluate their current capabilities, communicated them clearly, and maintained focus, adoption would accelerate.
We also need to adjust to a new computing paradigm. Using ChatGPT effectively requires understanding its conversational nature and need for iteration. Many users find it unsettling that computers might make mistakes and require human-like interaction - pointing out errors, seeking clarification, or asking for confidence levels. This approach feels counterintuitive after decades of using computers differently.
The challenge lies less in missing functionality and more in poor communication that breeds confusion and hesitation. ChatGPT's success (beyond being an excellent product) stems from its clear purpose. Unlike intrusive AI features that pop up offering to rewrite emails, it's a distinct tool with an obvious function. Google and Apple's strategy of embedding AI throughout their operating systems may actually deter users. A better approach might have been to focus on developing more capable Siri and Google Assistant, emphasizing the ability to converse naturally with devices. Because people are being pitched magical technology that’s just around the corner we’re not compelled to learn how to use what we have today.
It’s OK to be skeptical about any of this. A lot of this stuff is vaporware, a lot of it is also potentially dangerous. Prompts use up a ton of energy and hallucination can cause real harm and confusion. I’m not a die hard accelerationist and part of me wishes things went back to the way they were before ChatGPT dropped. But, I also believe this is transformative technology, right now, today. Downplaying it won’t change that.
ANONYMOUS BANKER
Any robot can make an S-1. Good ones do more.
A few weeks ago, David Solomon (CEO of Goldman Sachs) mentioned that his engineering team, consisting of thousands of engineers, used AI to draft an S-1 in just a few minutes. Typically, this process takes highly paid bankers and lawyers more than a week. This isn’t that interesting, and I’m surprised Solomon is boasting about tech that should push clients to ask for reduced fees; AI agents don’t require Brioni suits.
It’s no secret that content is a commodity in the banking world. Most of the material produced, including S-1s, is built on the back of cribbing historical work; we aren’t journalists, so plagiarism isn’t a sin. While it’s convenient that investment banks are building or buying tools to help junior staff organize, summarize, and publish material more efficiently, the resulting leverage could simply lead to more superfluous presentations, models, and analyses (Jevons paradox on crack).
Where AI truly has the potential to transform investment banking and investing is through the proliferation of agentic, autonomous tools that can drive real-world (“kinetic”) actions. For example, I know of a top-quartile venture firm that spends millions on structured and unstructured data feeds, funneling them into an LLM-powered database. This system automatically alerts the investment team when entrepreneurs or companies exhibit promising signals. It has already led to several new investments, and you can envision a future in which there’s effectively a “synthetic” deal partner, complete with AI engineers demanding a share of carry for finding top-tier opportunities. Remarkably, this technology only requires a team of fewer than 10 engineers to power it.
Looking ahead, the real value will come from tools that empower senior dealmakers and investors to source, engage, and nurture relationships that lead to new investments and generate fees. That’s where AI-driven leverage can meaningfully move the needle.
18 good links from our group chat
Perplexity wants to turn the SuperBowl into downloads with $1M giveaway
New resistance media lives on YouTube
How the Bullwark is building a media business in the information space.
Andreessen Horowitz makes a statement hire
Marine veteran Daniel Penny became a national figure after he used a chokehold on an agitated subway rider. He was cleared of criminally negligent homicide last year. He will join the firm’s American Dynamism as a deal partner, a practice that backs companies that support American interests.Competing with China and how we are losing crucial battle for cars
Ironic scamming is still scamming
Enron was a fun art project. Then they made a meme coin.Does fact checking work?
Despite rigorous fact-checking efforts, many still embrace misinformation. How useful is fact checking anyway?The state of the nation… prosperous but sad
The State of the Nation report dissects how we are doing across 37 measures.Just cause you can’t figure out the money use case doesn’t mean it can’t kill you.
Nobel Prize winner Geoffrey Hinton, the physicist known for his pioneering work in AI, told LBC's Andrew Marr that artificial intelligences had developed consciousness - and could one day take over the world.Revenge of the south.
Are Northern Europeaners getting their comeuppance. I doubt it.Broligarchs vs Populism
Despite his clashes with new factions emerging in the Republican Party, Bannon argues that Trump is still central to advancing a populist agenda. Good listen.How to make money in media with Kara Swisher.
Kara at her best.Is OpenAI just an app company now?
A good take on what happens when LLMs commoditize.
GOOD PRODUCT
The Swamp Princess is a master.
Douchii’s Tiny Desk Concert is a very Good Product. Glad she won the Grammy.