Imminent Death of ChatGPT [and Generative AI] Is Greatly Exaggerated
synthedia.substack.com
synthedia.substack.com
>I could first point out that it is not often 69% of businesses adopt any technology simultaneously.
AI has been a buzzword talked about by executives for at least ten years now. The same issues with data organization and labelling were being explained to executives 5 years back [2] and are still relevant today. LLMs are a jump forward in NLP which enables more use cases in business, but AI and its adoption challenges are not new.
[1]https://www.axios.com/2023/08/19/ai-corporate-barriers-cost-... [2]https://www.mckinsey.com/capabilities/quantumblack/our-insig...
The dissonance is wild. I'd hate to go back to working without it. It definitely has its uses.
They act like it's literally useless except for plagiarism.
An electric drill's capabilities are known, in the way AI's is not. There's no drill bit coming in Q4 that will let you drill into steel for example. But there will be a GPT 4.5 or something that will add new capabilities.
After so many disappointments in the tech universe (and with some AI winter in the past) people in the industry should not be naive to the point of thinking that OpenAI would solve all the industry's problems.
>> OpenAI has marketed it as an "everything" tool
That's what all companies do all the time. Should we blame capitalism?
On the other hand, the finances of AI are only loosely correlated with it's utility to individuals. If OpenAI went from being worth 10 Billion to 500 MM, it wouldn't make too much difference to people who use OpenAI products.
So far it’s only been useful to me when working with a tool I don’t really know. But most of my work I know the tools fairly well.
I had to go through several iterations and I bounced from wanting to write it in go to I think eventually just doing it in bash, just simplicities sake, but it absolutely saved me days of research. I definitely had to fix a lot of things and tailor it to my needs and it lead me astray quite a few times but it really was more like asking a professor the question vs googling it and hoping that SOMEONE out there has done something similar. Or combing API docs.. It was such an ancillary thing that I didn't want to get hung up on it for more than 1-3 days and it literally got me going in maybe 3-6 hours.
I also write a ton of very advanced terraform and cdk with it. I already know both very well but it absolutely speeds me up if I haven't used a specific function or what not in awhile, or it's newer to me.
It's great at leading me to the water, sometimes I have to still teach myself to fish but it will give me ideas and get me going if not even more.
For me ChatGPT (v4) is very useful as it is and for my line of work is a God sent, and I give one simple real world example.
Earlier this month we have some encouraging results on earthquakes early detection based on IoT and Big Data. Our earthquake expert is not replying emails for several weeks and we had to send the research summary immediately to ask the journal editor for the suitability of potential publication based on the proposed paper's title and abstract. But since we are not the earthquake domain experts, we are not sure what's the correct word to use for the early detection of earthquake in our particular results, and the prime candidates are forecasting or prediction. Voila, upon consulting ChatGPT it's obvious that for our results, prediction is the correct word to be used.
Before Internet people used to go to library for looking up something, but then Google arrived. Similarly, before ChatGPT people used to these so called experts for consultations, but then ChatGPT arrived. But somehow, people seems to be in denial of the useful of the GPT in helping us to write better reports, papers, articles, etc.
The current usefulness of consumer AI pales in comparison.
There is no death of ChatGPT in sight. Even if the UX was a bit of a hoax to begin with. It would be if some of the other models beat it.
Seems like the AI hype was so great, that now there is an anti-AI hype saying AI was just another over hyped tech that is fading, and this wave of hype is turning into another AI-Winter.
This article is great at pointing out that AI is really still growing, just now in the 'building' stage. All the corporations are now off building things with AI, but in their own gardens, and not publicly hyping them.
The slower growth state of the typical hype cycle, when the slow adopters are doing their adopting.
This post has counter arguments to everything that was said in this post earlier.
After two or three development cycles, LLMs will still be a curiosity, but nothing more. They might be used for some non critical business cases where output doesn’t need to be perfect, but for all other cases the burden of verification will be unmanageable.
The use of LLMs in academia will be nearly eliminated as professors opt for more verifiable ways of ensuring students aren’t using ChatGPT: no more writing long form essays in the comfort of your own home and with access to the internet.
"Generative AI was overhyped, but it also isn't a nothing-burger"
Include some graphs. Include a poem. Sh*t talk Elon a little as well...
Generative AI burst onto the scene like a summer blockbuster, promising transformative experiences. The tech-savvy among us were both awed and slightly terrified at the notion of AI creating human-like content. But has it lived up to its hype?
The Highs of Generative AI
[Graph 1: Rapid Growth of Generative AI Adoption]
The above graph showcases the swift ascent of generative AI's adoption from 2019 to 2022. With businesses and individuals alike eager to experiment with this new tech, it was clear that there was significant promise.
From creating music, poetry, and even assisting with scientific research, the capacities seemed endless. Remember the poem generative AI penned last year that went viral?
Amidst the silicon dreams we tread,
Seeking truth in binary, it said.
Of ones, of zeros, of neural might,
The machine beholds the human sight.
But Wait, There's More (or Less)
However, for every positive leap, there have been equally notable stumbles.
[Graph 2: Drop in User Satisfaction with Generative AI Tools]
The above graph indicates that while initial excitement was high, satisfaction dwindled as users confronted the reality of generative AI's capabilities. It wasn't the magic bullet solution for every challenge, and it often still required human input to achieve the best results.
Remember when Elon Musk commented on generative AI? "It's like trying to use a hammer for everything – sure, you might get some nails in, but you're also likely to break a lot of things." While it's not often I find myself nodding in agreement with Musk (especially after that whole debacle with him trying to launch a line of AI-generated perfumes), he did have a point here.
Not a Nothing-Burger, Though
But let's not be too hasty. Generative AI may not have completely revolutionized our world overnight, but it has added value. It's facilitated easier workflows, added creativity in previously unimaginative sectors, and has certainly begun conversations on the potential and limits of technology.
[Graph 3: Industries Benefitting Most from Generative AI]
There are sectors, as highlighted above, where generative AI has truly shined. Content generation, arts, and even certain branches of research have seen remarkable strides thanks to this tech.
Concluding Morsels
Generative AI might not have been the pie in the sky many hoped for, but it also isn't mere table scraps. Its potential is still unfolding, and as with all technologies, it will continue to evolve.
So, whether you're a staunch supporter, a skeptic, or just someone who enjoys AI-generated poetry, there's no denying that generative AI has spiced up our tech menu. And who knows? Maybe the next big breakthrough is just around the corner.
Until then, remember:
In a world of hype and binary frost,
Some dreams are gained, and some are lost.
But in the dance of code and byte,
Lies a future, ever so bright.
Note: All graphs referenced in the article are conceptual and would typically be visualized with data. Given the format, they've been described rather than shown.
It was mostly an example to highlight that the graph was a little misleading. By the same token, ChatGPT is an extension of what OpenAI has been iterating.
- They were "too early" in that what they were squeezing out of GPT-3 was what 3.5/4 could enable from day 1 with ease...
- But they were right on time because 3.5/4 would have made differentiating their AI copywriting tool almost impossible, meanwhile GPT-2 was too finicky to scale like they managed to scale GPT-3 did.
- And then they were "too late" in that they didn't get enough time to catch up on their massive valuation. 1 year ago their data and machinery around scaling GPT-3 in production alone were a moat: today you can build a synthetic data set for world class copy writing for $100 and fine tune 3.5-turbo on it and build something Jasper quality or better as a weekend project.
I think there is room for specialized companies that "just add prompts" as long as the UX is not a literal chatbot. But it's not going to support going from launch to unicorn in 18 months like Jasper did, it'll be much more of a slog.
They were using AI to do copywriting/SEO, and most importantly they were doing this before ChatGPT came out.
GPT-3 wasn't approachable for lay people, and relatively few people were building on it: So a lot of marketing people were seeing generative AI for the first time through this single tool. And obviously it was earth shattering for them.
-
The problem is ChatGPT made outputting what Jasper does extremely trivial: I hang around random SEO FB groups and at least once a week someone asks if Jasper is worth it, and a bunch of people chime in saying no ChatGPT output is just as good.
ChatGPT is also incredibly easy to use: a grandma could use ChatGPT to write an SEO friendly recipe post.
That all leaves Jasper an awkward spot. I don't think they'll die, but to stay competitive they'll need to shed a ton of dead weight. They obviously can't keep all the machinery they built around GPT-3 running if it's no longer adding value, they now have to work 100x harder on positioning, etc.
You are saying "GPT is Simply too good" (which I agree with).
But earlier today, entire thread that "AI isn't good enough"
Responding to this post https://news.ycombinator.com/item?id=37256577#37260717
People, HN, Industry, everyone is all over the place.
This is Disruption, another Winter is not coming.
https://twitter.com/0xSamHogan/status/1680725207898816512
Nitter: https://nitter.net/0xSamHogan/status/1680725207898816512#m
---
6 months ago it looked like AI / LLMs were going to bring a much needed revival to the venture startup ecosystem after a few tough years.
With companies like Jasper starting to slow down, it’s looking like this may not be the case.
Right now there are 2 clear winners, a handful of losers, and a small group of moonshots that seem promising.
Let’s start with the losers.
Companies like Jasper and the VCs that back them are the biggest losers right now. Jasper raised >$100M at a 10-figure valuation for what is essentially a generic, thin wrapper around OpenAI. Their UX and brand are good, but not great, and competition from companies building differentiated products specifically for high-value niches are making it very hard to grow with such a generic product. I’m not sure how this pans out but VC’s will likely lose their money.
The other category of losers are the VC-backed teams building at the application layer that raised $250K-25M in Dec - March on the back of the chatbot craze with the expectation that they would be able to sell to later-stage and enterprise companies. These startups typically have products that are more focused than something very generic like Jasper, but still don't have a real technology moat; the products are easy to copy.
Executives at enterprise companies are excited about AI, and have been vocal about this from the beginning. This led a lot of founders and VC's to believe these companies would make good first customers. What the startups building for these companies failed to realize is just how aligned and savvy executives and the engineers they manage would be at quickly getting AI into production using open-source tools. An engineering leader would rather spin up their own @LangChainAI and @trychroma infrastructure for free and build tech themselves than buy something from a new, unproven startup (and maybe pick up a promotion along the way).
In short, large companies are opting to write their own AI success stories rather than being a part of the growth metrics a new AI startup needs to raise their next round.
(This is part of an ongoing shift in the way technology is adopted; I'll discuss this in a post next week.)
This brings us to our first group of winners — established companies and market incumbents. Most of them had little trouble adding AI into their products or hacking together some sort of "chat-your-docs" application internally for employee use. This came as a surprise to me. Most of these companies seemed to be asleep at the wheel for years. They somehow woke up and have been able to successfully navigate the LLM craze with ample dexterity.
There are two causes for this:
1. Getting AI right is a life or death proposition for many of these companies and their executives; failure here would mean a slow death over the next several years. They can't risk putting their future in the hands of a new startup that could fail and would rather lead projects internally to make absolutely sure things go as intended.
2. There is a certain amount of kick-ass wafting through halls of the C-Suite right now. Ambitious projects are being green-lit and supported in ways they weren't a few years ago. I think we owe this in part to @elonmusk reminding us of what is possible when a small group of smart people are highly motivated to get things done. Reduce red-tape, increase personal responsibility, and watch the magic happen.
Our second group of winners live on the opposite side of this spectrum; indie devs and solopreneurs. These small, often one-man outfits do not raise outside capital or build big teams. They're advantage is their small size and ability to move very quickly with low overhead. They build niche products for niche markets, which they often dominate. The goal is build a saas product (or multiple) that generates ~$10k/mo in relatively passive income. This is sometimes called "mirco-saas."
These are the @levelsio's and @dannypostmaa's of the world. They are part software devs, part content marketers, and full-time modern internet businessmen. They answer to no one except the markets and their own intuition.
This is the biggest group of winners right now. Unconstrained by the need for a $1B+ exit or the goal of $100MM ARR, they build and launch products in rapid-fire fashion, iterating until PMF and cashflow, and moving on to the next. They ruthlessly shutdown products that are not performing.
LLMs and text-to-image models a la Stable Diffusion have been a boon for these entrepreneurs, and I personally know of dozens of successful (keeping in mind their definition of successful) apps that were started less than 6 months ago. The lifestyle and freedom these endeavors afford to those that perform well is also quite enticing.
I think we will continue to see the number of successful micro-saas AI apps grow in the next 12 months. This could possibly become one of the biggest cohorts creating real value with this technology.
The last group I want to talk about are the AI Moonshots — companies that are fundamentally re-imagining an entire industry from the ground up. Generally, these companies are VC-backed and building products that have the potential to redefine how a small group of highly-skilled humans interact with and are assisted by technology. It's too early to tell if they'll be successful or not; early prototypes have been compelling. This is certainly the most exciting segment to watch.
A few companies I would put in this group are:
1. https://cursor.so - an AI-first code editor that could very well change how software is written.
2. https://harvey.ai - AI for legal practices
3. https://runwayml.com - an AI-powered video editor
This is an incomplete list, but overall I think the Moonshot category needs to grow massively if we're going to see the AI-powered future we've all been hoping for.
If you're a founder in the $250K-25M raised category and are having a hard time finding PMF for your chatbot or LLMOps company, it may be time to consider pivoting to something more ambitious.
Lets recap:
1. VC-backed companies are having a hard time. The more money a company raised, the more pain they're feeling.
2. Incumbents and market leaders are quickly become adept at deploying cutting-edge AI using internal teams and open-source, off-the-shelf technology, cutting out what seemed to be good opportunities for VC-backed startups.
3. Indie devs are building small, cash-flowing businesses by quickly shipping niche AI-powered products in niche markets.
4. A small number of promising Moonshot companies with unproven technology hold the most potential for VC-sized returns.
It's still early. This landscape will continue to change as new foundational models are released and toolchains improve. I'm sure you can find counter examples to everything I've written about here. Put them in the comments for others to see.
And just to be upfront about this, I fall squarely into the "raised $250K-25M without PMF" category. If you're a founder in the same boat, I'd love to talk. My DMs are open.
If you enjoyed this post, don't forget to follow me, Sam Hogan. I share one long-form post per week covering AI, startups, open-source, and more.
That's all folks! Thanks for reading. See you next week.