⚙️ Neurosoft CEO on navigating the hype of BCI tech

Subscribe | Advertise

Good Morning. Today, at 10 A.M. PT (1 P.M. for my fellow east-coasters) OpenAI will be unveiling … something. There were rumors last week that the company was going to launch a search-enabled ChatGPT product on Monday, but Sam Altman himself said that this announcement has nothing to do with search OR GPT-5. So, we’ll see, and we’ll break it all down for you tomorrow. 

I also had a fascinating conversation with Dr. Nicholas Vachicouras, the CEO of Neurosoft, a brain-computer-interface startup. Read on for the full story.

In today’s newsletter:

• 💬 Updates are coming to ChatGPT today 

• ☣️ More OpenAI safety researchers are leaving

• 🪧 Pause AI organization protests creation of AGI

• 🩺 Neurosoft CEO on navigating the hype of BCI tech

Mysterious ChatGPT updates are coming

OpenAI confirmed that it will be going live later today (10 A.M. PT) to “demo some ChatGPT and GPT-4 updates,” a move that closely followed rumors that a search-enabled ChatGPT is coming.

• CEO Sam Altman, however, shot down those rumors last week, saying: “Not gpt-5, not a search engine, but we’ve been hard at work on some new stuff we think people will love! Feels like magic to me.”

Then on Friday, The Information reported that OpenAI is “preparing to demonstrate technology that talks to people — using sound as well as text — and recognizes objects and images.” This might be relevant to today’s big reveal.

• Around 40% of you said in our poll last week that a search-enabled ChatGPT will threaten Google’s hold on the market. But an almost equal number said it all depends on the details of the product.

  • The rest don’t see a way for prompting to replace the ease of a quick keyword search. 

More OpenAI safety researchers have left the building 

Two OpenAI safety and governance researchers – Daniel Kokotajlo and William Saunders – recently departed the company. 

Kokotajlo’s Less Wrong profile says that he “quit OpenAI due to losing confidence that it would behave responsibly around the time of AGI.”

• He added that he declined to sign a non-disclosure agreement with the company upon leaving to retain his ability to “criticize the company in the future.” 

• In doing so, he gave up a tremendous – the unknown – amount of equity. 

This isn’t what you want to see from a company that has made it their mission to develop artificial general intelligence. I feel it’s important to note that OpenAI got its start as a nonprofit research lab before deciding to start selling stuff. Now, it’s valued somewhere in the region of $80 billion. OpenAI did not respond to my request for comment.

Together with Brave 

Feed Your AI with Real Data

The Brave Search API brings affordable developer access to the Brave Search Index: an independent index of the Web with over 20 billion webpages. *

What makes it special?

• Built on real page visits from actual (anonymous) humans, which filters out tons of junk data.

• Independent and proprietary, meaning no Big Tech biases or extortionate prices.

• Refreshed daily, so it’s always accurate and up-to-date. 

The Brave Search API can be used to assemble a dataset to train your AI models, and help with retrieval augmentation at the time of inference. And it does so affordably.

Integrating the Brave Search API into your workflow translates to more ethical data sourcing and more human-representative datasets. 

Try the Brave Search API for free for up to 2,000 queries per month.

Second international protest to pause AGI development

On the same day as OpenAI’s mystery unveil, the Pause AI organization is holding its second international protest in 13 cities around the world, notably London, New York and San Francisco. 

• The goal is to convince people with the power to act that the creation of artificial general intelligence — OpenAI’s guiding star — isn’t worth the risks. 

Let’s talk about AGI:

I have spoken with numerous ethicists, researchers and computer scientists about AGI – there is a big concern among many that focusing on fears of an out-of-control AGI (which has been spread by the leaders of prominent AI organizations, including Sam Altman and Elon Musk) could eclipse attention on current harms from current systems, such as algorithmic discrimination, widespread disinformation and economic inequity. 

Many scientists have also been dismissive of the idea that AGI might ever be possible; Suresh Venkatasubramanian, an AI researcher who in 2021 served as a White House tech advisor, told me last year that the idea of AGI is little more than “religious fervor.”

• The point, according to Gideon Futerman, a member of the protests, is to stop corporations from attempting to develop AGI. “An AGI race cannot be won and should never be fought,” he said. 

💰AI Jobs Board:

• Research Engineer ML Efficiency: Netflix · United States · Los Gatos, CA · Full-time · (Apply here)

• Director of AI Innovation: X Boson · United States · Arizona · Full-time · (Apply here)

• Principle Researcher — Generative AI: Microsoft · United States · Remote · Full-time · (Apply here)

 🗺️ Events: *

• Ai4, the world’s largest gathering of artificial intelligence leaders in business, is coming to Las Vegas — August 12-14, 2024.

  • Don’t wait — passes are going fast. Apply today for a complimentary pass or register now for 41% off final prices.

📡 Competitions in AI & Machine Learning:

• Atmospheric Physics with AI: LEAP — Develop machine learning algorithms to accurately model atmospheric physics in an operational climate model. ($50,000 prize)

• Predict New Medicines: Leash Bio — Develop machine learning models to predict the binding affinity of small molecules to specific protein targets. ($50,000 prize)

• Develop AI Safety Benchmarks: SafeBench — Develop benchmarks related to Robustness, Monitoring, Alignment, or Safety Applications ($250,000 in prizes)

*Indicates a sponsored link

Together with Superhuman

Superhuman was already the fastest email experience in the world, designed to help customers fly through their inbox twice as fast as before.

We’ve taken it even further with Superhuman AI — an inbox that automatically summarizes email conversations, proactively fixes spelling errors, translates email into different languages and even automatically pre-drafts replies for you, all in your voice and tone.

Superhuman customers save 4 hours every week, and with Superhuman AI, they get an additional hour back to focus on what matters to them.

Get 1 month free of Superhuman

Neurosoft CEO on navigating the hype of BCI tech

The world of deep tech and AI is home to a bunch of even deeper offshoots. Among them is the fascinating industry of brain-computer interfaces (BCI). Though this kind of tech has been around for decades, it’s lately been gaining attention, a common result of what I like to call the ‘Elon Musk Effect.” (Musk, of course, started a BCI company called Neuralink, which is currently in the midst of its first human trial). 

But the idea of inserting a chip into the brain is an approach that Dr. Nicholas Vachicouras thinks is outdated. 

• Brain chips like Neuralink’s are penetrative, meaning they damage brain tissue as a trade-off of insertion. So, Vachicouras set out to create a different option – a flexible, soft electrode array that sits on top of the brain, rather than being inserted into it. 

Thus, Neurosoft was born. 

The combination of electrode hardware and AI-powered software allows the device to record, detect, stimulate and decode brain activity, a base from which Neurosoft is exploring a wide variety of medical applications. 

• Tinnitus – The area Vachicouras is most excited about – involves a less conventional BCI approach. In this instance, the Neurosoft device would stimulate the portion of the brain that generates that false ringing in a patient’s ears to reduce or eliminate it. This is Vachicouras’ clinical North Star for the company. 

• Epilepsy – Neurosoft is close to beginning an epilepsy trial, in which its electrodes will guide brain surgeons to help them determine which parts of an epileptic patient’s brain to remove. 

• Synapsuit – One of the biggest promises of BCI involves the control of robotic prostheses, which could enable paralyzed people or amputees to function again. Neurosoft is in the early stages of a partnership where its device will be used to record, decode and reconstruct — using AI — brain activity related to physical movement, which can then interface with such devices. 

“We’ve been working on these electrodes for many, many years,” Vachicouras told me. “We now have all the safety data that we need.” 

Demystifying BCI:

The positive of the Elon Musk Effect is that the BCI field has more eyes on it than before, and more eyes mean more investors. The negative is that Musk’s hyped-up statements about BCI tech — that it could allow people to experience alternate realities, for example — has, for a lot of people, eclipsed the actual science. 

“I still think it's a net positive, but it has its downsides,” Vachicouras told me. 

• Myth: More electrodes and sensors are better than less.  

• Reality: “What matters is clinical impact,” Vachicouras said. “There's no proof today that if you do 1,000 electrodes versus 100, you’re going to have better outcomes.”

• Myth: The software is more important than the hardware. 

• Reality: Even with AI, Vachicouras said, “the reason why this is not widespread today is a hardware problem.” For a device to properly interface between a machine and the wet environment of the brain requires overcoming biological, chemical, mechanical and electrical complications. 

• Myth: We can read & control minds with brain implants. 

• Reality: The mind, as J.K. Rowling wrote in Harry Potter and the Order of the Phoenix, is “not a book, to be opened at will and examined at leisure.” 

  • Vachicouras said that the biggest misconception he’s encountered is simple: people just don’t know that scientists do not know how the brain works. 

  • “We definitely know more today than we knew 10 years ago, or even three years ago, and I think it's advancing very, very fast,” he said. “But it's still a lot less than what people think.”

• Pythia: An AI tool designed to assist founders.

• MeetGeek: A tool to record and transcribe meetings.

• Translate.video: A tool to transcribe videos into 75+ languages.

Have cool resources or tools to share? Submit a tool or reach us by replying to this email (or DM us on Twitter).

*Indicates a sponsored link

SPONSOR THIS NEWSLETTER

The Deep View is currently one of the world’s fastest-growing newsletters, adding thousands of AI enthusiasts a week to our incredible family of over 200,000! Our readers work at top companies like Apple, Meta, OpenAI, Google, Microsoft and many more.

If you want to share your company or product with fellow AI enthusiasts before we’re fully booked, reserve an ad slot here.

One last thing👇

That's a wrap for now! We hope you enjoyed today’s newsletter :)

We appreciate your continued support! We'll catch you in the next edition 👋

-Ian Krietzberg, Editor-in-Chief, The Deep View