AI doesn’t solve the real problems. It’s surface level pseudo productivity.
The only new things or features I’ve really seen are related to AI. So it’s just AI adding AI to stuff. I’m still waiting for this big renaissance of innovation.
If a startup can be throw together quickly with an LLM, where is the moat? Anything they make could be ripped off by a big player overnight.
If the cost to build drops, it stands to reason that the valuation drops as well, unless they can actually get traction in the market.
I’m not sure if you’ve watch the show Shark Tank, but Kevin brings this idea up a lot. If they have a generic product and no protection around it, it can be easier and cheaper to just spin up a clone with their own team than to take on the risk of an investment.
The market seems flooded with slop startups. I’ve seen many where people use AI to make a site to pitch the idea, but there isn’t even a product. It’s all smoke and mirrors. The industry is full of opportunistic charlatans right now. A good crash is needed to clear the field before real growth can happen.
Those who are building products with AI, probably aren’t advertising the fact. I’ve seen some stuff in the App Store tho I assume nothing of at first, then read the description and it was clearly written by AI, which makes me assume the whole thing was vibe coded. The stuff is out there, but it’s all so small and spread out, that all noise and no signal.
I could be way off, but that’s my read on the situation.
I'm advocating for a mass hiring event and huge growth push, not throwing together MVPs with LLMs!
Big dck money. Big dck ROI -- The next great Batman film, a World of Warcraft 2, autonomous robots on the moon, one of those humanoid robots doing my laundry!
The cost to build has dropped considerably, yes, the barrier to entry has lowered, but the bar is raised and valuations have skyrocketed and set records with trillion+ dollar IPOs too. More capital flowing into individual companies than ever before, albeit a small set of companies.
The potential has gone up, 10-fold, 100-fold.
To me, the market is saturated with pickaxe vendors looking for scrap metal miners, that is, model wrappers hoping to be the next wave of SaaS serving a largely misunderstood - disenfranchised (laid off from their jobs, no hope for the future) - customer base (gen y/z developers) that isn't meeting overall sales expectations. Nobody likes AI. Nobody needs it. They're also being told "Hey, don't be developers anymore". It's very confusing messaging!
So you got Google accidentally dominating most of it anyway just by being in everyone's house and pocket. Deer in headlights they're like "Who me?" They weren't even trying. Gemini surprises everyone. Surpasses everyone. Haha. Don't worry, they're doing a lot worse now that they are trying.
I digress, beyond "AI", what is needed is clearly... <AlexCarp>The Application Layer</AlexCarp>
> Those who are building products with AI, probably aren’t advertising the fact.
Perfectly fine - and gets to the spirit of what I'm talking about. Forget AI, focus on the product (and customer).
This all assumes the code was the bottleneck. That rarely seems to be the case. I work in a large organization full of bureaucracy. Getting people to agree on what to build, and sorting out dependencies between team, and general communication doesn’t seem to be helped by AI, and those are the bottlenecks. I had the CIO tell my boss project X was the most important thing, just get it done. I ignored all the meetings, made every decision myself, and was done in 1-2 weeks (without AI). Then the bureaucracy entered. Over the next 2 years I was on daily meetings, architects would randomly show up and tell me this or that needed to be changed (always for the worse, imo). Hell, we spent 4+ months debating a list of ports to open. I wanted to throw myself out the window. I started out with a list of ports that were actually required, we could sub in an AI answer for that, and we still spent half a year tinkering. “We”… them, I just had to sit there and listen to it.
Will a great Batman film be written by AI? It seems like it would be a boring and predictable script without and soul, rehashing old themes that have all been done. And the script is only a minor piece of the puzzle. A lot of moving parts have to come together to make a great film. Does AI have taste? Good taste?
If AI can’t solve the bottlenecks, it’s not actually going to speed up delivery.
Good question. Big companies participate in incrementalism. Many companies invent one big thing and ride it for decades. Few build new things and succeed. (e.g. Apple). Some try, e.g. Microsoft with Xbox, Meta with VR, but even then often it ends up being billions of lost profits. Innovation often come from small groups of persistent and intelligent people and their first few acquisitions as they succeed. e.g. Microsoft with DOS, Windows and Office. They're still riding it.
The last startup boom began in the mid-2000's (around when Y Combinator began, you can read all about it on Paul's blog), when cloud computing was still new and lowered the cost of creating online businesses. It was also in the wake of the tech bust - a lot of tech employees became unemployed.
My view is, when the layoffs complete, the dust settles and you have millions of tech workers who are willing to accept only equity, with a new technology they can rent for $200 a month from DeepSeek or Anthropic or OpenAI, we may see the green shoots of a new wave of disruption.
Maybe there's a bit more to a company and its software than the quantity of code it generates? And maybe there's a built in moat that a lot of these companies have that takes a lot of effort (and marketing, etc) to overcome?
People don't just jump social networks because a new one exist, there has to be a compelling reason to migrate over to it. And for a lot of people, there needs to be a sufficient number of people (and probably existing friends) already using the app for them to expend the effort to start using it.
It's not like there haven't been other attempts at competitors for these over the years, and you haven't heard about them probably because none of them ever got any traction. And that takes more than just building software.
Like DuckDuckGo has been competing with Google search for years (it first came out back in 2008, it looks like), and likely is way better developed than anything someone will crank out with A.I. in a short period of time. And it still has a tiny fraction of Google's marketshare.
I'm not saying there won't eventually be other competitors to these, but it'll take more than just whipping up an MVP with A.I. in a few weeks.
AI can help you write code, but it still can't help you get customers, or attention. In fact, it makes it worse. AI helps with introducing more noise, more content, and more people clamoring for attention. It's a lemon market.
Mainly am asking why those with attention (or VCs who can get it) are not introducing more noise, more content, and more clamoring - as you put it.
I thought that was the whole idea?
Lemon market - maybe - but it remains to be seen IMO. The market seems almost entirely unrealized. There are only a few major AI companies in the media, and it seems like innovation is largely directionless or stalled.
I haven't really seen any "startup boom" for AI yet. It's all the same kind of AI wrapper startup still. The most innovative thing in San Francisco right now is AI complaince.... Compliance... I'll leave you with that.
'those with attention' is a unicorn. Look at Facebook + Google - they own attention. Yet a huge % of products they make are duds. It is much much harder than you think to get lots of people to use your product.
2) Making a billion dollar company is too expensive. Netflix lost money for years. Amazon lost money for years. Tesla lost money for a decade. Spotify took 18 years to become profitable. The AI companies are losing billions of dollars each month.
Agree LLMs are more like a calculator than the person who uses it. You actually still need the person to use it! And you want that anyway for accountability and ownership. That's one of the big reasons I think companies should be hoarding AI talent right now.
And 2 - yeah I really mean "product focused" companies, not these YC AI wrapper replicas you see everywhere (no offense to YC but hey, they know what they are doing lol). Those companies must be only there to drive usage to the model providers they wrap.
The disruptive potential is not limited to those who use AI, or those who offer it in their product to end users. The potential is here regardless. Because of the fact workers use it in their workflows - in Adobe software, in their IDE, and so-on, it's effectively here.
The company does not need to do anything special - just hire.
It's simple: AI can't really write software. AI is good in generating garbage code averaged from millions of repos.
Here is an example of software that can scale with a single person assuming AI can write software: a web-based alternative to Photoshop built on wasm. Figma is analogous to that and it's worth $10Bn.
Assume for a moment that the code of cost is zero, and the barrier to entry for coding is gone. Furthermore, let's say you don't want to do anything innovative, you just want to clone, so a huge part of your product management burden goes away.
The playing field is now completely level with any and all comers who want to attack Adobe. You now not only have one huge competitor, but the potential for hundreds of other competitors.
How are you going to convince someone that your execution is the one that's going to displace Adobe and steal market share? Why would I invest in your variant rather than one of your competitors?
The logical conclusion of this is that even if Adobe released its source and the theoretical price drops to zero, you're finding yourself at best in the same situation as Linux distro companies in 1999.
Assuming perfect LLM code generation (a huge assumption that is nowhere close to reality), a company built on this is going to replay the same open source dynamics, with additional steps/expense.
Photopea predates LLM coding by a lot. The fact that it was made by single developer is not particular, so was Figma.
The reality is, Photopea is unique because no one with LLMs can easily produce something similar. Also Photopea does not compare to Adobe offering similar to how Figma does. Not even close.
AI is not just disrupting companies and contributing for layoffs but it's also changing the buyers and customers.
Why build a new Google when the potential Google customers now just get their results from LLMs instead of SERPs?
How to be the next Jobs when there's no actual new product, just essentially an assistant tool?
How to create the next Disney, Netflix, Blizzard or Pixar when LLM's are still not good enough to produce quality content and many people are also starting to despise AI generated content? Art is still an intrinsic Human thing, even if LLMs are able to produce it.
So, although these LLMs are super amazing and definitely revolutionary, this whole AI/AGI hype is still a bit on bubble territory imho so I'm not really sure we'll see what you're hinting at with this generation of LLMs.
Also, wasn't AI supposed to be the last product we'd ever need to invent and then it's utopia and singing kumbaya for all eternity?
We won't need new companies, big personalities or heroes once Skynet/Hal9000/GLaDOS/Agent Smith/etc are in charge :)
There are opportunities to build a new Google without AI. In that case AI creates an opportunity in an unexpected way. If Google goes all-in on AI, the opportunity is wide open for someone else.
And we do see that (Proton, DuckDuckGo, even iCloud is becoming very Google-like in recent years).
As for "AI is not good enough" I remain unconvinced - I've seen production quality work, granted not 1-shot from a model but produced by a talented artist.
I never said AGI, (I believe that's only a marketing term from OpenAI). The other stuff at the end of your comment has nothing to do with what I'm talking about here! I really am focused on the low cost of code, video, and other media compared to the massive opportunity to serve content to people worldwide and how much money and disruptive potential is being left on table.
Can you frame your question in a single sentence? Even after reading everything you wrote above, I’m not clear what the essence of your question really is.
But here's another thought: Code and pixels are still cheaper than ever before.
Another one: The earning potential from code and pixels is the same or greater than it was before AI.
Put it together. If it was real estate where land was cheap but selling at the same rate as before, then investors would be all over it.
Innovation is something you have to believe in - I get that - but from a purely finance perspective why is there not a massive movement to obtain swaths of low cost, high potential IP? Data, engagement, emails and cards on file. Series A-F growth, greater fools and IPOs - either way it's massively generative of productive wealth, and you'll get a few game changers here and there.
I fail to see why it's not a great opportunity for, say, a $1B+ fund in just about any category.
The only new things or features I’ve really seen are related to AI. So it’s just AI adding AI to stuff. I’m still waiting for this big renaissance of innovation.