AI is the infrastructure, NOT the product!

“OpenAI just killed my startup!”

You hear it time and time again, more recently after Apples WWDC announcements and their hilariously marketed “Apple Intelligence”.

So does this mean we all face the same risk of “CompanyX just killed my startup” moments, or is there hope for the various AI founders out there?

Our Braiv Journey…so far

From my own personal experience founding & building Braiv, and having COUNTLESS conversations with both VCs, prospective customers, and friends (including fellow founders), this sentiment seems widely true.

“Your startup will be eclipsed by LLMs”

“Wait until GPT 5 comes out…you’ll see”

“Your product will end up being a free solution by OpenAI”

Wow…what a way to view AI startups…that a simple enhancement to an existing model will “kill” your idea. The idea that has brought you to building a platform dedicated to solving a real world problem. Will ChatGPT just kill it? Really?!

Well…

I DISAGREE…but only if you have this mentality: “AI is the INFRASTRUCTURE, not the product”.

Erm…”what you talking ‘bout Ben?!”

The Realistic Current State of AI Startups

Well ironically, I think Sam Altman of OpenAI actually said it best:

There’s those who look at AI models or LLMs and think, “Cool I can enhance this for my users”…and guess what, the next iteration of that LLM provides that same feature…your startup just died.

But there’s the small few that look beyond where the landscape is now, and tries to envisage a more futurist approach to things…they build something that will only get better when “GPT 5 comes out”.

When your Solution is Providing AI as the Product

If you, as the unknown startup founder, are trying to build a solution built upon single AI solution, you are in the 99.9% of startups that are due to fail…in my opinion. Examples to this may be:

• Text-To-Speech
• Video subtitle generation
• Image generation
• Animation generation

Now…I deliberately said that 99.9% of startups focussing on AI as the product will fail because there are the anomalies that are proving their worth, in all the above categories. But they are indeed the anomalies.

The reality is, most of these productised AI solutions are just pretty wrappers and refinements to existing open source AI models that basically any techo can spin up and leverage. Let’s use the above list and see what they are hiding under the hood:

• Text-To-Speech = Tortoise TSS & RVC
• Subtitle Generation = OpenAI Whisper
• Image Generation = Stable Diffusion
• Animation Generation = Erm…also Stable Diffusion

There’s no use hiding it…you’ve just reskinned or refined something that will inherently become more accessible and commoditised by the open source community.

How Should AI Founders Approach Things from Here?

Look…I don’t have all the answers to this…but here is my approach:

NEVER marry yourself to a single AI model for a single purpose, nor isolate your platform to providing a single AI-enabled solution.

Truly, the future of this space for “the little guy” is to treat AI in the same way you may treat the hugely accessible cloud infrastructure layers now available from the “big dawgs” such as Amazon, Google or Microsoft.

We are WAY beyond the days of founders building a “cloud storage and sync solution” (think Dropbox or Box). Why is this? Well because the infrastructure layers to build these solutions are so widely available now, it's completely pointless. But that doesn’t stop founders using these technologies as valuable integrations within their own platforms…that serve a broader purpose, and solve “real” problems.

AI is no different…am I getting my message across?

How Braiv is Building on the “AI as the Infrastructure” Mentality

Braiv sees all the enhancements in AI as the perfect infrastructure to build an integrated platform for what is, and has long been, one of the more complex solutions in market…VIDEO.

We recognised that sure, organisations can:

• get their videos captioned using Veed.io
• translate the captions using any number of translation solutions
• even dub video content using ElevenLabs
• finally they can use Vimeo to deliver individual dubbed videos.

Cool…that’s 4 platforms to do what are some darn simple actions…all in the name of creating a more engaging and accessible video experience for their viewers.

Well at Braiv we want to integrate all of these “features” (or better put “AI infrastructures”), into a single, automated platform that allows users to get all the captions, translations and dubbed content from a single workflow…but better yet, deliver this content to their audiences with one of the most innovative online video players in the market: Braiv Player.

Imagine it…the vision is simple:

1. Upload/import your video content
2. Set a preset workflow to that video upload
3. Get the most accurate video caption files generated
4. Get all your captions translated with the best translation engine
5. Get your videos dubbed with the best quality voices
6. Deliver the video in a gorgeous, branded video player.

From points 3–5, Braiv will be integrating at least 7 of the best-of-breed AI models available.

“Your platform will be eclipsed by LLMs!!!!”

No…no it won’t…as LLMs improve, so will our platform.

In fact by leveraging the best of the best LLMs in market, our platform will inherently be the best of the best for our customers use case. Why?

BECAUSE AI IS THE INFRASTRUCTURE, NOT THE PRODUCT!

Conclusion

I’m not a writer…clearly…but hopefully my point is clear.

Don’t marry your solution to single AI functionality, or even just a single AI model enhancement.

The future of AI is seamless integration of the best-of-breed models to solve real world problems.

At Braiv, we are building our platform with this very ethos in mind, and as such, we believe that our startup will NEVER be “killed by OpenAI”.

Thanks for reading. If you’re interested in what we are building, please sign up!

Much love…Ben from Braiv.

‍