NineTwoThree AI Studio

How to be AI-first without actually being AI-first: - Buy the .ai domain for your business - Read 17 different AI newsletters every week - Sign up for ChatGPT, Claude, Perplexity, and Gemini - Also sign up for a random startup that is a wrapper for the above 4 LLMs - Never use any of the models besides ChatGPT - Put “AI expert” in your LinkedIn header - Use ChatGPT to write a LinkedIn post or two, then realize it’s really bad at that - Add a Jira card at your job to “Decide on our AI strategy” (forever in TODO state) Congrats, you’re AI-first without actually doing anything! Now, how do you ACTUALLY become AI-first? Understand where AI can meet the needs of tomorrow’s entrepreneurs.

We have a few openings to build AI models for companies starting in April. Here’s why you should jump at the opportunity: We have dozens of clients that have returned their AI investment with us within the first year. - We helped a consultancy firm go from 500k → 1M with a 200k Investment in the AI Model. - We helped reduce maintenance orders by 90% for a massive mechanical shop creating 8 figures of savings. - We created an entire new million-dollar AI product for a startup, and helped hire their entire data science team to run it. Now they are getting Series B interest. Our product managers are MIT / Harvard?/ MBA trained, and they only need 5 weeks to roadmap your AI project. We are 1/10th the cost of *Big Consulting Firm.* (We've competed and won bids for clients against them.) All of our AI projects have been put into production in less than 5 months. It feels to us that these companies are finding a slot machine that they know will hit 7 - 7 - 7 this year. Our process takes them from idea through execution and into production. By the time we are training their team on how the AI model works the company is seeing massive returns. There are only a few slots - are you next? (And if you recommend someone, we offer 5 figure cash payouts for the introduction.)

This is a reminder that OpenAI has a family of tools, not just ChatGPT ??

Our team stopped using ChatGPT overnight, without a single complaint. And we didn't lose a step. And this trend will continue until ChatGPT releases hardware or reverse partners with Microsoft. Yes, OpenAI has built the best model, and it's still the best model we use for all of our clients today under the hood. (In fact, we are turning 6 figure investments into 7 figure returns for dozens of clients.) But ChatGPT, the app, is no longer used by our team. We got rid of it with a single announcement in our general Slack channel. Why? We found something better for our use case. Google released an LLM that works securely inside our Google Workspace, and is 80% as good as ChatGPT. - Finally, we can prompt questions about our clients, securely, in an LLM (we couldn't before.) - Finally, we can copy transcripts from phone calls into Gemini’s AI studio and ask it to summarize what the call was actually about while remaining as secure as an email. - Finally, we can take a product call that I missed, run the transcript into AI studio, and extract the purpose, meaning, and 5 takeaways without worrying about data leakage. We can search through emails faster, find data in our drive from a semantic reference, or search previous proposals to get ideas on the paragraph we are stuck on.

A lot of industry leaders are worried about adopting AI - they're feeling the pains of the cloud shift that took 5x too long. But AI is here to stay. It won't take as long as you think to get on board, and you'll be ahead of competitors (who are starting to look at AI already) Invest in your future!

"Wow, that new AI model performed better in one irrelevant edge case, we'd better switch"

An AI privacy wish list, if you will.

Sometimes, it’s really hard to tell which AI tool you should be using. 3.5 Sonnet vs o1 vs Llama vs Gemini vs whatever else…what do these words mean? Here’s a handy guide. (All of these tools have free tiers)

There will only be two types of companies in the future: those that use AI, and those that are out of business.

Your 3rd grader can do something that ChatGPT can’t. Try something. Ask ChatGPT to draw someone writing with their left hand. It can’t. Seriously, if you can get it to do it, show us! “Wait, I just heard we’re 13 seconds away from AGI!!!!!” GPT, right now, isn’t “learning” anything. It’s drawing on its training data. And GPT’s training data: - Saw millions of photos of people drawing with their right hand - Doesn't actually think through what it’s trying to accomplish. - Forgets its mission after finding what it thought was the answer. And until we see AI start learning to do this, we’re not even close to AGI.