🔥 Sam Altman under fire

ALSO: Create sound effects and AI agents

Estimated Read Time: 5 minutes

Sam Altman can’t stop smearing his own reputation. After weeks of hot news, former board members are speaking out. Plus new automation tools, AI generated sound effects, and robotaxis. Let’s dive in!

• 🔥 Why Sam Altman was fired last year.

• 🎧 ElevenLabs can generate sound effects now.

• 🚕 China’s cities open new robotaxi routes.

• 🦾 Anthropic releases new automation “agents”.

• 🎨 Canva’s lets everyone be a designer.

🔥 Why Sam Altman was fired last year

What happened: Helen Toner, a former OpenAI board member who helped oust Sam Altman last year, has broken her silence during a podcast about why Altman was fired.

The details: 

• She claims Altman constantly withheld important information from the board.

• This is especially egregious since the board was designed to operate separately from the company and oversee their commitment to safety and AI’s effect on humanity. They were not profit-driven.

• She claims the board was not informed about ChatGPT’s release on November 2022. They found out through Twitter. 😆

• Altman did not disclose his ownership of the OpenAI startup fund to the board.

• The board received concerning reports from various executives about Altman’s behavior in October.

• Altman was described as creating a toxic atmosphere and engaging in psychological abuse.

• Overall, Altman consistently lied to the board or created excuses or misrepresentations in order to achieve his own goals.

Why it matters: This is just nail in the picture. Using all of the previous knowledge we have, it’s clear that Sam Altman is driven by progress and profit and will do whatever it takes to achieve his goals at the expense of his employees and overall safety. But… that won’t matter to ChatGPT users. All we really need to know is whether or not the product is good. His direction is very much the direction of AI overall.

💰 China has launched its third state-backed investment fund, known as “Big Fund III,” to strengthen its semiconductor industry and reduce reliance on foreign technology. Basically they’re after full self sufficiency in chip making. It’s pretty successful so far at 344 billion yuan (~$47.5 billion), which is larger than of their any previous funds. Btw the US and EU have their own funds at $280B for the U.S. CHIPS Act and €43 billion for the EU Chips Act. Nobody wants to rely on another country for AI chip manufacturing… the risks are too great.

🤓 GPT-4 is smarter than humans at complex reasoning, according to a new study. This level of reasoning is important for social intelligence and predicting behavior. The short and sweet is: AI is getting better at higher order functions, including social awareness and cognitive thinking. It’s less impressive when you realize how much faster computers think than humans… and they still answer questions wrong. It’s also concerning, as the paper notes, because AI is also getting smart enough to manipulate people (if it wanted to).

🗣️ Amazon is rolling out AI voice search to Fire TVs. You can ask open ended questions about shows or movies and the Fire TV will attempt to answer them. It’ll also recommend more shows and movies (of course). A small QOL feature.

🎧 ElevenLabs can generate sound effects now

What happened: ElevenLabs just released a tool which allows you to create sound effects from a simple prompt.

The details:

• The tool is available to all users starting today and works by typing in simple prompts like, “Wood creaking,” “metal clanging,” “cop siren,” etc.

• The tool can create instrumental loops up to 22 seconds long, such as guitar riffs and trumpets.

• Free users can create about 60 sound effects per month, or 10,000 character generations (about 150 per sound byte).

• The model was legally trained on Shutterstock’s library. No copyright issues here (supposedly).

Why it matters: This is pretty big. We now have voices, music, and sound effects generated by AI. Soon we’ll have video. Eventually anybody with some creativity and know-how will be capable of creating high quality productions from scratch with a low budget.

🚕 China’s cities open new robotaxi routes

What happened: Beijing has begun testing robotaxis with human staff inside, covering a 40-minute route between Beijing South Railway Station and the suburb of Yizhuang.

The details: 

• The new route connects Beijing South Railway Station with Yizhuang, a suburb.

• The intent is to be fully driverless by early next year. Right now they still need human staff in the car for emergencies. (Good luck.)

• The city plans to extend robotaxi services to Beijing Capital International Airport. A useful idea for many cities.

• The goal is to operate 1,000 robotaxis in Beijing by the end of 2025.

Why this matters: This will someday be commonplace in most cities and rural routes. For now it’s in early stages. But these driverless cars will be a reality and China isn’t messing around. Beijing could be an example of how to integrate robotaxis into modern life. (Poor Uber drivers though…)

🦾 Anthropic releases new automation features

What happened: Anthropic just released “tool use” or “function calling” which allows you to hook up Claude 3 to an external API to complete tasks automatically.

The details:

• Examples of automated tasks include creating a bot to purchase commerce products and an email assistant.

• This is not a full AI agent… but it’s close.

• The tool can access any external API and analyze data to act accordingly, such as providing customer support.

• Pricing is based on text volume processed, with 1,000 tokens equating to about 750 words.

• The tool has already been available in beta, with positive feedback from testers.

Takeaway: Functions like these will be the backbone of AI automation in the workplace, enterprise, and in personal life. Google I/O revealed similar features, but Anthropic’s tool is available right now. I’d suggest giving it a spin to get a head start on creating AI agents.

🎨 Canva’s lets everyone be a designer

What happened: Canva recently held their “Create” event where “work redesigned” became their mantra. With that came a host of new AI features enabling anybody, regardless of skill level, to create designs.

Some of the features:

• Magic Design creates ads and presentations quickly.

• Magic Grab allows you to grab any element in a photo (like your ex 👀) and move them around.

• Magic Write is copywriting assistant who writes in your own voice using written samples.

• Enhance Voice improves audio quality. (Similar to Adobe’s offering.)

• Magic Media creates videos from simple prompts. 🧐

• Canva Marketplace: There’s now a marketplace for AI-powered apps which integrate with Canva.

Takeaway: Canva is central to many professional’s workflow. Yes, none of these tools are revolutionary or special. They’re available elsewhere. But their availability in a program used by over 180 million users is a stepping stone in how we use AI everyday. It’s easy to forget that simple QOL tools are some of the best features AI has to offer.

 💧 Thirst quencher?

Source: Midjourney

Give us your feedback!

Got feedback? Rate today’s newsletter by clicking below! 

Got more personal feedback? Fill in the box when you rate us and give us your criticism and feedback.

Thank you for reading! 

❤️Share the (ARTIFICIAL) Love!

Got a friend who wants to appear as smart as you? An AI fanatic or someone shaking in their boots of getting replaced? Share us by clicking the link below and you’ll get our love ;)