😺 Ex-OpenAI Employees speak out

Welcome, humans.

We've teamed up with Clipt to create some insanely cool video edits for our podcast. The result is somewhat akin to how a cat feels after a hit of catnip: utterly euphoric, completely enthralled, and irresistibly hooked.

We think you’ll love the revamped look. Take a peek at yesterday’s pod, where Pete breaks down Nvidia's journey, what GPUs and intelligence factories are, and Nvidia’s new AI chips.

Here’s what you need to know about AI today:

• Former + current OpenAI/DeepMind employees sign an open letter on AI safety.
• Helen Toner claims Sam Altman lied about OpenAI’s safety processes.
• Companies are shifting from building with one AI model to multiple.
• All the major chatbots including ChatGPT had outages yesterday.

ex-OpenAI employees speak out against the company’s commitment to safety.

If you've noticed a spike in "AI safety" headlines, you're not alone.

Here’s the scoop: A small group of people in Silicon Valley, namely at OpenAI, Google, Anthropic, and Meta, is building AI, potentially the most powerful and consequential tech in human history.

And with great power comes great responsibility. These companies claim their focus is on building AI that’s safe, not harmful. But at the end of the day, they’re all companies, and companies are designed to prioritize making $$$.

Recently, there’s been growing discontent among some in the AI field that “safety culture and processes have taken a backseat to shiny products.”

In plain speak: profit > safety.

• The quote above is verbatim from Jan Leike, who co-led OpenAI’s now-disbanded Superalignment team, which was formed to prevent AI systems from “going rogue.”
• You might also recall Ilya Sutskever, OpenAI’s co-founder and the other Superalignement co-lead who was involved in the decision to oust Sam Altman last November (Ilya departed OpenAI last month).
  • Another former board member, Helen Toner, disclosed last week that they ousted Sam partly because he provided false info about OpenAI’s safety procedures “on multiple occasions.”

Now, even more insiders are speaking up.

Yesterday, 13 current and former OpenAI/DeepMind employees published an open letter on AI safety called “A Right to Warn about Advanced Artificial Intelligence.”

The TLDR: There’s no way to hold AI companies accountable for their safety pledges.

Governments are slow, and employees can’t report dangerous but non-criminal AI activity to the public (OpenAI recently retracted a clause that pressured employees to sign NDAs or risk losing their vested equity).

Granted, these are just 13 individuals among thousands developing AI, and they no longer work at these orgs. More coming on this topic soon!

Hear Pete’s extended take on the political battles inside OpenAI, safety vs. capability, and superalignment’s death in last month’s podcast (Apple Podcasts, Spotify, & YouTube).

FROM OUR PARTNERS

Hire a world-class AI team trusted by Fortune 500 companies and leading startups.

Building AI software in your company is hard. You’ve got three options:

1. Track down technical engineers and pay them a fortune.
2. Try to DIY, but building AI products is highly technical.
3. Use AE Studio.

Their stacked team of AI experts builds the perfect AI solution for your business. That’s why they’re the go-to for leading startups and Fortune 500 companies.

Their development, design, and data science teams work closely with your team to make sure your AI does exactly what you need it to.

Book a FREE consultation today!

Around the Horn.

• ChatGPT, Gemini, Claude, and Perplexity all faced outages yesterday.
• Shutterstock made $104M in revenue last year solely by licensing their images and videos to AI companies.
• The proportion of Kruze Consulting clients using multiple AI models rather than just one jumped from 1% last summer to 15% in April.

Treats To Try.

1. *The AI Conference brings together the brightest minds in AI. Get $200 off your ticket for the two-day event in SF on Sept. 10-11 using code “Neuron24”.
2. GetWhy is a research tool that extracts insights from video interviews using AI (raised $34.5M).
3. Cohere builds AI models specialized for enterprises and just raised another $450M.
4. True Fit is an online shopping tool that helps consumers determine how clothes will fit.
5. Vantage Discovery helps online retailers build search engines that help customers find the right products (raised $16M).

*This is sponsored content. Advertise in The Neuron here.

A Cat's Commentary.

Where do you #Neuron?!

Send us a picture of where you read The Neuron here (with some spicy commentary), and you might get featured next week!

Get two friends who’ll wake up for the sunrise at 5:30am and read The Neuron with you! spoiler: taken from Noah’s iPhone 13 ;)

Paul (from Verona, Italy) writes: “From MA, but on vacation in Italy and still need my daily fix of AI news!”

Matthieu (from Pokhara) writes: “Trying to match the ‘The Neuron’ font color of the day from my hotel room.”

From St. Louis: “WFB = working from bakery.”