AI and The Trough of Despair

I have been mulling over a concept. It’s called the Trough of Despair (or disillusionment). If you look back at the history of technology, there are patterns you can observe, and I think we are living through one right now. In a similar way that economies have boom and bust cycles, there is also a similar pattern for new technology. It has been studied a lot, and it even has a name. It’s called the Gartner Hype Cycle. There are 5 stages:

1. The Technology Trigger – The Technology Trigger happens when a new technology explodes onto the stage. There is a buzz, including news, press releases, murmurs on YouTube, demos at conferences. There are no applications yet, but there is a tangible hum that is growing.

2. The Peak of Inflated Expectations – The next stage is the peak of the hype cycle, where this technology explodes and you hear about it everywhere. You can’t avoid it. Everyone is talking about how it will change everything. New products and applications start releasing every day, and you can’t keep up! Everything looks polished and promises the world. However, if you look closer, these products tend to over-promise and under-deliver. There are hundreds of new companies springing up as they head to the gold rush.

3. The Trough of Disillusionment – The third phase is when the technology fails to live up to its promises. People start second guessing. Investors get nervous and slow down. People react with skepticism to new product demos. Adoption is slowed and goes negative. Companies start going under. The news stops covering it.

4. The Slope of Enlightenment – The fourth phase is when the slow adoption and slow improvement of the tech begins. It’s slow, but it is real. The product is refined. The use cases are narrowed, but become more effective. The Slope of Enlightenment is where the benefits of the technology become understood and accepted.

5. The Plateau of Productivity – Finally, the tech becomes mainstream and sees widespread adoption but it’s not ultimately as disruptive as the hype cycle said it would be.

The “trough of despair” describes a phase in the lifecycle of a hyped technology where, after an initial surge of excitement and lofty promises, it fails to fully deliver on its transformative potential. This happens when a technology achieves about 90% of its promised capabilities but struggles to bridge the final 10% needed for widespread disruption. The gap leads to disillusionment, reduced investment, and skepticism, before eventual refinement and adoption.

Here’s what happens: new technologies explode onto the scene with bold claims of revolutionizing industries, improving lives, reshaping economies, and disrupting everything in sight. If you’re observant, you can spot this cycle by noticing phrases like “game changing,” “revolutionary,” “This will replace every job,” “this makes (insert industry here) irrelevant,” “this is as bad as it will ever be.” Sound familiar? I bet you’ve seen some clickbait like this about AI lately!

This creates a “hype cycle,” as described by Gartner, where expectations peak during the “inflated expectations” phase. AI has promised to kill off the entire VFX industry for god’s sake, and within the year!

This brings us to the next phase. When the technology falls short—due to technical limitations, scalability issues, or unforeseen complexities—it enters the trough of despair, where enthusiasm wanes, and critics highlight its shortcomings. This happens when the technology, which quickly hit 90% efficiency, struggles to get to 100%. Hallucinations anyone?

This is a critical juncture. Technologies that enter this zone now enter a track that has two possible outcomes. Some technologies stagnate or fail entirely, unable to overcome their limitations, and they become obsolete. Others persist in this trough for years, finding limited use while struggling to achieve mainstream adoption.

A few eventually emerge through persistent iteration, solving the “last 10%” of challenges—whether that is reliability, affordability, or user acceptance—to achieve significant impact.

Let’s quickly look at a handful of examples, and you can guess if they made it through the trough or not.

In the early 1990s, virtual reality was heralded as the future of entertainment, gaming, and even social interaction. Companies like Sega and Nintendo announced VR headsets, and this fueled public imagination with visions of immersive digital worlds. The technology promised to transport users into fully interactive 3D environments, disrupting everything from gaming to education.

The Bubble: By 1995, VR systems like the Nintendo Virtual Boy and many more hit the market, backed by massive media hype. Expectations soared as developers promised seamless, photorealistic experiences.

The Trough: The reality was far less glamorous. Early VR systems suffered from low-resolution displays, clunky headsets, limited field-of-view, and motion sickness. The Virtual Boy, for instance, used a red-and-black monochromatic display that caused headaches and failed to deliver immersive experiences. Processing power was insufficient, and the systems were super expensive, often costing thousands of dollars. The final 10%—delivering comfortable, affordable, and truly immersive experiences—was unattainable with the era’s technology. By the late 1990s, VR faded from mainstream attention.

Autonomous vehicles have been touted as a game-changer for transportation, promising safer roads, reduced traffic, and mobility for all. Companies like Tesla, Waymo, and Uber invested heavily, with predictions of fully autonomous cars by the late 2010s. Sirens were sounded in the media about the imminent collapse of all jobs in the taxi industry and the trucking industry.

The Bubble: In the mid-2010s, demos of self-driving cars navigating controlled environments fueled optimism. Elon Musk claimed Tesla’s vehicles would achieve full autonomy by 2018. Venture capital and public markets valued autonomous driving startups at billions, expecting rapid deployment.

The Trough: The final 10%—achieving reliable autonomy in complex, real-world conditions—hasn’t quite worked out yet. Unpredictable pedestrian behavior, adverse weather, and regulatory hurdles remain unsolved. Public trust nose-dived after high-profile accidents, such as Uber’s 2018 fatal crash. As of 2025, fully autonomous vehicles remain in a prolonged trough with widespread adoption likely years or even decades away.

In the late 2000s, 3D television was marketed as the next evolution of home entertainment, promising cinema-like immersive experiences in living rooms. Major manufacturers like Sony, Samsung, and Panasonic invested heavily, and films like Avatar (2009) amplified consumer interest in 3D technology.

The Bubble: By 2010, 3D TVs were prominently featured at consumer electronics shows, with broadcasters like ESPN launching 3D channels. The technology promised to transform how people watched movies, sports, and gaming, with predictions that 3D would become the standard for TVs.

The Trough: The final 10%—delivering a seamless, accessible, and enjoyable 3D experience—proved insurmountable. 3D TVs required special glasses, which many viewers found cumbersome and uncomfortable. Content was limited, and producing 3D programming was costly. By the mid-2010s, consumer interest waned, and manufacturers like LG and Sony discontinued 3D TV production. The technology failed to emerge from the trough, effectively dying as a mainstream product.

The convergence of AI-generated art and cryptocurrency-backed non-fungible tokens (NFTs) in 2021 sparked a frenzy, with claims that these technologies would revolutionize the art world. NFTs promised verifiable ownership and monetization on blockchain platforms like OpenSea. High-profile sales, such as Beeple’s $69 million NFT artwork, fueled visions of a new digital art economy.

The Bubble: In 2021, the NFT market exploded, with trading volumes reaching $25 billion. Crypto art was hailed as a democratizing force for artists and collectors. Investors and celebrities poured money into NFT projects, expecting exponential growth and mainstream adoption.

The Trough: The final 10%—establishing lasting value, accessibility, and cultural acceptance—proved elusive. The NFT market crashed in 2022–2023, with trading volumes dropping over 90% and many projects losing most of their value due to speculation, scams, and lack of intrinsic utility. High transaction fees on blockchain platforms and environmental concerns about energy-intensive crypto mining further alienated users. As of 2025, AI NFTs linger in a prolonged trough with most projects struggling to regain relevance or financial viability.

The trough of despair underscores the challenges of translating technological promise into reality. The last 10% often requires not just technical breakthroughs but also infrastructure, regulatory frameworks, and user acceptance.

New technology screams from the mountain tops, “This time is different! This will change everything in 6 months!” However, transformative technologies often take decades to mature, if they succeed at all. By understanding the trough of despair, we can better navigate the information flooding at us.

So, is generative AI for motion design in the trough of despair? Yes, I believe so. I would guess that generative AI for motion design is either entering or already in the trough of despair, but with a strong likelihood of surviving it in some form. Here’s why:

Generative AI tools like Runway ML, Adobe Firefly, Kling, Midjourney etc. have been hyped as game-changers for motion design, promising to automate complex tasks like keyframing, rotoscoping, and video generation from text prompts. The buzz around AI art was sky-high, with claims of revolutionizing creative workflows.

But recent reports suggest a cooling-off period. For example, Gartner notes that generative AI broadly has slid into the trough due to mismatched expectations, weak ROI, and challenges in scaling practical applications. Motion design-specific tools face massive issues: while they can generate impressive visuals, the outputs often require significant human cleanup to meet professional standards, and features like lip-sync or complex animations can feel “wonky” or unpolished.

There is something off about the results. The physics is a bit off. The humans are a bit waxy. The movements are a bit too jerky or a bit too smooth. Nothing quite feels right to you on first glance. This is the 90% problem. It’s 90% there, but this gap between promise and delivery screams that we are entering trough territory.

There is a change in the wind. It reeks of unrealized promises. How many times will we hear people say things like “it’s not quite ready for production, but it’s close” before we decide it’s never actually going to be ready?

The generative AI boom is fizzling. Back in 2023, headlines screamed about AI revolutionizing economies, with promises of boosting global GDP by 7%. Remember when AI was supposed to churn out Oscar-worthy animations or viral motion graphics? Yeah, that hasn’t happened.

A recent report points out that 30% of generative AI projects are being scrapped before they even leave the prototype stage. See here. In motion design, tools like Runway ML can churn out quick video drafts, but designers often spend hours fixing clunky transitions or unnatural movements, negating the time-saving hype.

Tools like Kling AI can generate slick visuals, but they’re often derivative, lacking the emotional depth or originality humans bring to motion design. AI hasn’t delivered a single iconic song, film, or animation. The problem? It’s great at mimicking patterns but terrible at the nuanced storytelling that defines great art. As the industry cools on AI hype, executives are now questioning whether the tech’s worth the energy bills and ethical headaches. See here.

Other aspects of the tech are hitting a plateau as well. Scaling up models like GPT-4.5 barely improves accuracy despite ballooning expenses. Some of the updates are actually causing more hallucinations, not less! Add to that the ethical mess—copyright issues, biases in outputs, climate issues—and it’s no wonder businesses are pulling the plug. Generative AI’s not dead, but it’s stuck in a rut, far from the creative savior it was billed as.

What happens next? It’s anybody’s guess. I believe AI will most likely survive and be with us forever. I just think the hyperbole has reached max levels, and the reality is setting in.

Art is not something you can shortcut. There are no easy buttons. Art is about so much more. It’s part of being a human. It’s our experiences. It’s the process, which is not easy.

For me, the trough of disillusionment is a giant relief. I’m glad AI won’t take our jobs, at least not for a while. That gives me more time to do what I love. And that is a beautiful thing.

What do you think? Any insight would be great. You can reach me at joren@thepixellab.net

Good luck with your projects!