If you’re one of the 2,000 new subscribers from this past week…welcome! Thanks for joining and happy to have you here!

If you joined on accident and do not wish to be on this list, please feel free to hit the unsubscribe at the bottom of the page.

One of the biggest areas that will be impacted by AI over the coming years is investing. I’ve been using AI-driven trading bots for the last several years to invest my money. I typically make around 10% monthly returns from it.

If you want to learn more about them…I recently went through this course that was super helpful for getting started in just a few short minutes. Check it out here.

_{Read Time: 10 Minutes}

Today’s Menu

• Small Model, Big Impact—Is Size No Longer the AI Metric?

• Fed's Indecision and Middle East Tensions Rattle Markets

TODAY IN AI

Google Goes Open-Source?

Google just introduced Gemma 2 2B, a lightweight AI model with only 2.6 billion parameters that outperforms much larger models like GPT-3.5 and Mixtral 8x7B on key benchmarks. Trained on a massive 2 trillion token dataset, Gemma 2 scores 1130 on the LMSYS Chatbot Arena, matching or beating much larger models. It also delivers strong performance on MMLU and MBPP benchmarks, exceeding its predecessor by over 10%.

This model is open-source, signaling Google’s push to lead the emerging market for efficient, on-device AI solutions that don't compromise on performance.

Mark Zuckerberg has been an outspoken advocate of building open-source AI programs (using it in the new LLaMA 3.1 model). Zucks idea is that open-source AI programs foster more innovation among developers, reduce overall costs, and create a more accessible technological future for all.

However, a lot of key names in AI disagree and take the opposite route. OpenAI’s more recent models (anything beyond GPT-3) are closed source. Anthropic’s Claude LLM is closed source.

My two cents on this topic (as a small business owner who doesn’t have a huge team or billions of dollars in VC funding behind me) is that open-source is beneficial to a lot of smaller companies that don’t have the resources of major names in AI (like Meta, Microsoft, OpenAI, Google, Anthropic…etc). Only time will tell though…

^{Today In Crypto}

Fed's Indecision and Middle East Tensions Rattle Markets

Crypto prices plunged after the Fed's decision to leave interest rates unchanged, with Bitcoin dropping to around $64,500, down 2.5% in 24 hours. Broader digital assets fell nearly 4.24%, as measured by the CoinDesk 20 Index. Investors were spooked by cautious comments from Fed Chair Jerome Powell on potential rate cuts in September, compounded by escalating tensions in the Middle East after Iran vowed retaliation against Israel.

Riot Platforms reported a second-quarter loss tripling to over $84 million, exacerbated by Bitcoin’s halving, which slashed mining rewards and production.

Meanwhile, Ripple is allocating $10 million to tokenized U.S. Treasury bills on the XRP Ledger, reflecting the growing trend of tokenizing traditional assets for more efficient management and distribution in the crypto space.

Based on opinions from a few of my favorite anonymous Twitter accounts…

It seems like a lot of people in crypto feel that we’ve reached an interesting point of separation.

The BTC ETF came and went. The price went up a decent amount (high 40s into low 70s). But we didn’t see the CRAZY price increases that a lot of people expected (I.E. $150,000/BTC range).

The ETH ETF was just approved by the SEC about 10 days ago. No major price changes have occurred.

My gut feeling is that a lot of it has to do with the election. Trump is positioning himself as the “pro-crypto” choice.

I personally think we’re going to have a few stressful months of basically nothing happening followed by major price movement early next year (which direction? tough to say).

^{Today In Robotics}

Standard Bots Raises $63M To Bring “Cobot” Arms to Market

Standard Bots recently raised their Series B round led by General Catalyst. Standard Bots is mainly focused on the manufacturing space, providing robots that can handle dangerous, mundane, and repetitive tasks.

One of the main things that separates Standard Bots is their vision programming.

Rather than using coding to teach their robots how to do tasks, users can simply perform the tasks they want the robots to learn in front of the robot. The robots observe and replicate.

They’re already being used by companies like NASA.

I would keep an eye on this company in the future - looks like they’re doing a lot of interesting stuff.

Want to share The Future of Disruptive Tech with a friend, family member, or coworker? Pass this link along to them.

Thanks for reading!

Until next time,

Damon James