AI the Product vs AI the Feature

Author Avatar

Marques Brownlee

Joined: Mar 2024
Spread the love

AI the Product vs AI the Feature


The new Siri vs the RabbitR1 and Humane pin

Rabbit R1: https://youtu.be/ddTV12hErTc?si=tLR_GSXyRFtpgpJb
Humane AI pin: https://youtu.be/TitZV6k8zfA?si=vI4mZMhNvcGsOaXr

MKBHD Merch: http://shop.MKBHD.com

Tech I’m using right now: https://www.amazon.com/shop/MKBHD

Intro Track: Logic https://logic.lnk.to/DejaVu
Playlist of MKBHD Intro music:…

source

Reviews

0 %

User Score

0 ratings
Rate This

Sharing

Leave your comment

Your email address will not be published. Required fields are marked *

42 Comments

  1. That's the whole issue with OpenAI and already established AI startups and likely why Sam Altman wanted to push regulations to push out any competition with higher barriers of entry. All the existing smaller AI companies will be absorbed into the established giants. So, Google, Microsoft, and Apple, now Nividia will continue running the entire tech space until we actually do something about oligarchies.

  2. Some pretty good points are made here but I think one important thing this video ignores is that AI software for very narrow use cases can actually exist on its own as a "product" and not become a "feature" — in fact this has been the case for years in things like medicine and scientific research

  3. Definitely a "feature" – always has been, always will be.

    There's just no way we will carry more than one/two devices that need constant charging (notebook/phone/tablet + hearing device).

    The other distinction in AI that I'd like people to make is between consumer AI and enterprise AI. The hype tends to conflate these two when in reality, there's a huge gulf between them, in terms of implementation and application.

    Looking at the enterprise side, how will ordinary companies implement – not use, I think we know enough about the use cases – but actually implement AI?

    Will they all train their own Foundation Model? I doubt it, considering the costs. What I call AI-IaaS.
    Will they choose Foundation Model-as-a-Service (FMaaS)? In some cases, yes. A kind of AI-PaaS.
    But I'm betting the majority will realise (after spending millions in failed POC/Pilots) that they're NOT in the AI business – they just want to consume it, so they will opt for an AI-SaaS solution.

    No one is talking (much) about this and there isn't even a broad consensus.

  4. Love how a short and quick insight about a certain thought related to technology, you still make a video about it with the proper intro, editing etc, raising the value of your expertise in the industry. Even I enjoy watching this as it brought me closer to subscribing you as a human, rather than just a reviewer. Feels good to watch video from tech channel but is not limited to doing reviews about a certain tech product. Keep it up!

  5. well llama 3 is fucking slow on a 4090, so wont run local models on a phone haha. if laptop cards ever get 64gb VRAM then you can run local models, or wait for apple to get ddr6 ram in 2028

  6. Wow, I hadn't realized that a lot of products become features. Building that new feature into an existing product makes it more desirable as a product. Interesting to think about…

  7. I think the key question is, what is the barrier to new entrants integrating any emerging technology into their platforms. If the barrier to entry is low, it is a "feature". If the barrier to entry is high, it is a unique product. For GenAI, the barrier to entry is very low.

  8. AI needs some base to work on, like you need to give ai context of email to generate you email or make it better, they always need a product to work, so AI will always be a feature, it can't operate on it's own. Like best fuel for your car is still a feature, you can't do anything with just fuel.

  9. AI is so massive in possibility that it will change technology and every industry. Features are obvious, but when the singularity of ai happens. The Ai will be the product, as well as anything that it creates.

  10. Hey so maybe Siri will finally understand "never play this song again" on Music+ stations I've made, because no matter how many times I tell it that when a particular song I dislike comes on, it'll repeat an hour later. 😡 "Never play this song again" used to work but not anymore.

  11. AI is an algorithmic paradigm. It allows building software that has a better integration with the concrete world outside. Chatbots and LLM just managed to showcase the potential of artificial neural networks but they enable software that would otherwise cost too much to make with canonical deterministic algorithms.