5 Types of AI Company - Explained!!
HTML-код
- Опубликовано: 17 июн 2024
- In this video, we talk about FACES Framework to understand different types of AI companies.
This is my personal learning from different types of AI companies from foundation AI companies to GPT wrappers.
🔗 Links 🔗
❤️ If you want to support the channel ❤️
Support here:
Patreon - / 1littlecoder
Ko-Fi - ko-fi.com/1littlecoder
🧭 Follow me on 🧭
Twitter - / 1littlecoder
Linkedin - / amrrs 
Наука
Im watching this because im making a RAG using OLLAMA and local LLM, its quite challenging but im sure will be worth it
Love the concept for this video. Thank you!
Thank you, Glad you found it useful. I've been exploring deep concepts like this and I picked this one as there are lot of misconceptions!
Mistral, Cohere, SAP, META, Anthropic, Google, OpenAI, Alibaba.. some middle eastern companies. OH yeah Hugging Face watched a excellent talk on how they built a Code LLM.
00:01 Different types of AI companies explained
01:57 Foundational model innovators are R&D companies building large AI models.
06:10 Different types of AI companies explained
08:06 Companies fine-tune existing AI models for their own use cases.
12:25 AI companies are heavily burning VC money in competitive space
14:41 Be cautious about joining AI companies with heavy offers
18:29 Building a solid infrastructure is essential for managing AI models efficiently
20:18 AI companies need builders for sustainable software systems.
24:06 Understanding Standalone and Integrated AI Products
26:11 Starting an AI company with a specific problem
29:47 Building AI applications for niche markets adds actual value
31:30 Building application layer is crucial for AI companies
35:05 Different types of AI companies explained
Bless you. I think it is an accurate and useful hierarchy. The growth areas for most engineers is adaptive fine-tuning and general open source model wrangling ...LoRA for sure, but also adding vision transformers, creating mixtures, etc.
Loved the video. Made me think in a lot of ways I normally wouldn’t
Your videos are always insightful
love it , where is the end part companies list website??
What is your advice to one undergrad learning GenAI specifically to get placed?
Do lots of projects. Write about it. Repeat!
Any tip for experienced software engineer who looking for change in AI and ML space which type of company and role should be looking for. What skills should develop before we start applying?
Self learned basic model fine tuning using hugging face and still learning…
Really nice video ❤
May Allah bless you brother, you are realy delivering with your content
So nice of you. Thanks brother!
Really a nice explanation. May Allah bless you, brother
Building out infrastructure and integrating AI/mlops into production is probably the most important part. A lot of people can test llms and configure test environments. Placing in existing production is different story.
Great Video as usual! BTW What's on your left hand?
It looks like a henna tattoo, a temporary marking from a ceremony or celebration (e.g., from Eid or a wedding).
really liked the video
More helpful thanks❤
Nice informative video.
Very valuable content shared in this video
Thanks!
Glad it was helpful!
Gold Video! Thanks for compilation of the ideas. But, I might say that I am a bit distracted because of your left hand.
Jamba - by AI21 labs based on Mamba structure
Isn't AI wrapper and standalone Product company one and the same thing?
great video valuable time❤
plz plz don't stop making this kind of videos , make more these type of videos
bhai best video
very interesting video.
Can you make more video on standalone Ai for react , mern developers
As most of standalone Ai RUclips video just talk about voiceflow, stack ai no code tools platform
I want the US government to collect all private data of its citizens through OpenAI 🇺🇸, and the US government should completely take over SOTA models to do e-acc 🚀
❤
I am surprised u did not mention Anthropic as Foundation Model Innovators
Bro, can you please do a video on liquid neural networks in detail😅
🎯 Key points for quick navigation:
00:00 *🚀 Introduction and Purpose*
- Overview of the video's intent and audience,
- Introduction to the "FACES" framework for categorizing AI companies.
01:37 *📚 Explanation of the FACES Framework*
- Breakdown of the FACES acronym,
- Introduction to each category within the framework.
02:19 *🧠 Foundational Model Innovators*
- Description of companies building base AI models,
- Examples like OpenAI and Google DeepMind,
- Emphasis on the R&D focus and resource intensity.
06:46 *🔧 Adaptive Fine Tuners*
- Companies that adapt foundational models for specific uses,
- Examples of companies like Salesforce,
- Explanation of the fine-tuning process and necessary resources.
11:53 *🌐 Convenient API Providers*
- Companies offering APIs for easy model deployment,
- Discussion of market competition and business models,
- Examples include Replicate and Fireworks.
17:31 *🏗️ Essential Infrastructure Builders*
- Importance of infrastructure for AI deployment,
- Examples of companies providing tools for monitoring and deployment,
- Discussion of the essential nature of this category in AI ecosystems.
21:04 *🛠️ Essential Infrastructure Builders*
- Importance of infrastructure in AI product development,
- Requires more software engineers than AI engineers,
- Includes roles like kernel programmers and solution architects.
24:33 *🧩 Standalone and Integrated AI Products*
- Types of AI products: standalone and integrated,
- Examples of integrated products like Notion AI,
- Examples of standalone products like Jasper for copywriting.
26:49 *💡 Opportunities in AI Product Development*
- Building niche AI products for specific domains,
- Importance of identifying specific problems and solutions,
- Potential for quick profits with low initial investment.
30:33 *🏆 Market and Investment Dynamics*
- VCs' skepticism and market competition,
- The importance of unique selling propositions (USPs) in AI products,
- Examples of companies and their market positioning.
34:15 *📝 Summary and Final Thoughts*
- Recap of the FACES framework,
- Importance of different layers in the AI ecosystem,
- Encouragement to innovate in various AI product domains.
Made with HARPA AI
hugging faces , atari , nvidia
I guess those are api providers
6:04
Other than OpenAI? BLOOM (BigScience Large Open-science Open-access Multilingual Language Model) Although not a "company", more of a consortium. Its 176B open source model (1.7T token) is the basis for Bloomberg's closed source BloombergGPT Finance AI model.
BigScience is part of Hugging Face, isn't it?
@@1littlecoder The French government funded it, Hugging Face helped manage the initiative. After thinking about it, would you position Bloomberg as an Adaptive Fine Tuner? But like our organization, we use an open source base, and then build the stack (you called it a pyramid) of other components - which are also open source toolsets - in order to output some functionalities for Internal use ONLY. Our _Internal_ AI "stack" is justified by _Enhancing_ the existing core business of the company.
Like BloombergGPT is trained on *Bloomberg's highly proprietary data Lake* spanning back 50+ years, and can only be used internally for employees to help provide value. Bloomberg would never expose their data to any external provider on any level. If a small company uses an AI service provider to set appointments for example, there is no 'Barrier to entry' for their competitors. All small companies will use a common AI appointment setting provider, and very quickly, none of the organizations in the domain have any competitive value other than price. Basically AI will commodify all small businesses, just like Uber has commodified the transportation industry. ALL of the drivers - who are ALL small businesses - have no unique value, and that's why they will always be poor and poverty stricken. AI presents the same future for ALL but the largest enterprises...at least that's the way I see it.
@@1littlecoder The French government funded it, Hugging Face helped manage the initiative. After thinking about it, would you position Bloomberg as an Adaptive Fine Tuner? But like our organization, we use an open source base, and then build the stack (you called it a pyramid) of other components - which are also open source toolsets - in order to output some functionalities for Internal use ONLY. Our _Internal_ AI "stack" is justified by _Enhancing_ the existing core business of the company.
Bro I must ask, what's wrong with your hand?
I think it’s Mehandi 🤷♂️
My guess is he recently went to a wedding and it is a henna tattoo… Or it may be related to Eid (which was June 16 or 17 this year depending on where you are).
Diffusion models still have issues with hands 😂
@@mshonle I thought That too, but there's no pattern, he looks like he was slashed
@@rahulspoudelI can confirm that I was the guy who was sleeping in front of pc when this video was generating,
Nice informative video.