Good morning. Yesterday, OpenAI started rolling out its controversial GPT-4o voice mode feature to a small number of ChatGPT Plus users. OpenAI said the feature will roll out to all Plus users in the Fall.

The company said that the lengthy gap between demo and launch was due to extensive safety testing, specifically involving “100+ external red teamers across 45 languages.”

Let us know if you have access.

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

In today’s newsletter:

• 🚜 AI for Good: Smart precision farming

• 📱 Now you can make your own AI on Instagram

• 💰 Microsoft shares fall despite strong earnings beat

• 🌌 AI, AGI and the economy of Star Wars: ‘Humanity rises above the machine’

AI for Good: Smart precision farming 

The agricultural industry is a major water user, consuming roughly 70% of available freshwater resources. 

For years, though, tech has been working alongside the industry to increase sustainable efficiencies. 

The details: In 2024, the AI in agriculture market exceeded $2 billion; researchers expect it to exceed $5 billion before the end of the decade. 

• The root of it all is rapid data analysis; drones and remote sensors, combined with machine learning algorithms, can provide farmers with information and insights on soil moisture, pest populations and crop growth. 

• This enables farmers to more efficiently deploy resources to enable the survival and growth of their crops. 

Why it matters: Researchers have found that “with proper irrigation management through effective monitoring and optimal control, water can be saved, as well as providing a reduction in other indirect costs incurred from energy use in the form of electricity or fossil fuel for pumping, for optimal cost-effectiveness.”

Still, a careful cost-benefit analysis must be kept in mind here; if the water/energy use of running these models exceeds that which is saved, it probably isn’t as sustainably impactful as it seems. 

Get immediate access to up to 8 NVIDIA® GPUs, along with CPU resources, storage, and additional services through our user-friendly self-service console.

Learn more here today!

Now you can make your own AI on Instagram 

Meta this week announced that it is beginning to roll out AI Studio — a place for people to “create, share and discover AIs to chat with” — in the U.S. 

And so begins that circular future of non-human communication, wherein one person’s AI talks to another and reports back. Fun. 

The details: The studio, which is built on Llama 3.1, allows anyone to create AI characters, which they can then use “as extensions of themselves to reach more fans.” 

• Meta said that creators can create an AI character to respond to DMs for them, to reach more fans faster. 

• Meta added that any such responses will be clearly labeled to ensure transparency. 

Meta’s AI Studio policy states that it’s against the rules for users to instruct their AI to disseminate hate speech, identify their AI as a licensed professional of some kind or instruct their AI to help them engage in illegal activities, including child sexual exploitation.  

This implies that the system is capable of all of these things. 

Meta at the same time scrapped its celebrity AI chatbots, according to The Information, which the company launched last year.

• Canva acquires AI startup Leonardo.ai.

• Credo AI, a leader in AI governance software, announced that it had raised $21 million in new capital.

• Google-parent Alphabet's partnership with AI firm Anthropic under investigation in the UK (CNBC).

• AI bots talk dirty so OnlyFans stars don’t have to (Reuters).

• How a Microsoft App is Powering Employee Surveillance (404 Media).

• Perplexity is cutting checks to publishers following plagiarism accusations (The Verge).

• Meta to pay $1.4 billion to settle Texas facial recognition data lawsuit (Reuters).

Microsoft shares fall despite strong earnings beat 

When Microsoft reported earnings Tuesday afternoon, the numbers looked good. 

The company reported revenue of $64.73 billion, above expectations of $64.39 billion, and earnings of $2.95 per share, also above expectations. 

• Still, Microsoft shares tanked in after-hours trading following the release of the report, falling roughly 6%. 

• Part of the culprit for this might have to do with Microsoft’s capital expenditure number, which came in at $13.8 billion for the quarter, 5% above expectations. It might also have to do with Microsoft’s cloud revenue, which, at $28.52 billion, came in below expectations of $36.8 billion. 

The market’s reaction to Microsoft’s earnings beat here comes as investors are beginning to move away from these AI-happy, Big Tech giants which are trading at enormous price-to-earnings ratios. It highlights the import (and volatility) of investments in AI, which is running up huge bills but has yet to drive revenues to match. 

Deepwater tech analyst Gene Munster expects AI stocks to be down today on Microsoft’s cloud miss, but believes the “AI trade is intact.”

Global X analyst Tejas Dessai said that “questionable ROI around AI investments in the short term is weighing on the stock.”

The only AI Crash Course you need to master 20+ AI tools, multiple hacks & prompting techniques to work faster & more efficiently.

Just 3 hours — and you’ll become a pro at automating your workflow and save up to 16 hours a week.

Get the crash course here.

This course on AI has been taken by 1 Million people across the globe, who have been able to:

• Build No-code apps using UI-ZARD in minutes

• Write & launch ads for your business (no experience needed + you save on cost)

• Create solid content for 7 platforms with voice command & level up your social media

• And 10 more insane hacks & tools that you’re going to LOVE!

Register & save your seat now (100 free seats only) 🎁

AI, AGI and the economy of Star Wars: ‘Humanity rises above the machine’ 

As it stands today, artificial general intelligence (AGI) is a fiction. And often, it seems that the groups trying so hard to make it real are heavily inspired by fiction (Sam Altman and Her, anyone?) So, I thought it would be fun and interesting to look at one example of fiction as a lens to examine what AGI might mean if it were achieved here on Earth. 

A long time ago, in a galaxy far, far away …. 

Although a unified definition of AGI remains lacking, the droids of Star Wars have arguably achieved it. Both R2-D2 and C-3PO, for example, are able to communicate between different alien species, robots and computers without making mistakes (AKA hallucinations). They are reliable in their tasks, they can apply their knowledge to other scenarios and they even display emotion and what seems to be sentience (now, we don’t know the training data, but they seem to check off all the criteria on Gary Marcus’ list).

• But in Star Wars, humans aren’t out of a job. Han Solo still captains the Millennium Falcon. Human soldiers are still trained to fight AI-enabled robots (and other humans). Jobs seem plentiful in every field, from bounty hunting to healthcare, politics and investment banking. People work with robots, but people certainly still work. 

• In Star Wars, AGI is one thing. Human intelligence is something else entirely. And that unique capacity of human intelligence — boldness, creativity, ingenuity, spirit — often leads the way to victories that robots couldn’t have achieved (“never tell me the odds”.)

Here on Earth — despite the lack of scientific evidence that AGI will ever be possible — there is a belief among many that if AGI is achieved, it would surpass human intelligence and replace most labor, necessitating a Universal Basic Income. 

But AGI might not be as dramatically transformative as Elon Musk and Altman would have you believe. 

I sat down with Dr. Noah Giansiracusa, a math professor and AI expert, to discuss the thought experiment. He found it “very plausible” that achieving AGI is one thing; replicating/replacing humans with AGI is a much different equation. 

“Achieving doesn't mean universal deployment, it doesn't just happen even in our imagination,” he said. “There's just so much focus on achieving the timeline, but I feel like that's only part of the equation.”

• Part of this involves the cost. AI isn’t — and AGI wouldn’t be — magic. We’re talking about an incredibly resource-intensive technology, in finite materials, energy and money. And as resource-intensive as today’s large language models are, it seems likely that an AGI would drain significantly more resources, making its viability anything but a sure thing.

• Giansiracusa said there would likely be an economic calculus to the deployment of AGI or more powerful AI systems; he could see a world where it might simply be cheaper to employ humans than to use a hypothetical AGI. 

“Just assuming that something is capable and literally possible, doesn’t mean that all of a sudden everything just completely changes,” Giansiracusa said. 

He also envisions the potential economic impact of a cultural resistance to AGI “because we just want to live in a world of real things and real people.” Today, for instance, people pay more for handmade clothes, gourmet food and organic produce; in a world with AGI, people might well pay more for non-AGI, human-centric experiences and products.

The view that AGI will automate everything of economic value has “such a hyper-capitalistic feel that makes it feel like life is just a race toward efficiency and production,” Giansiracusa said. 

• “What are we really trying to do in life and in the world and society? And I feel like economics captures some of it, maybe a lot of it, but not all of it,” he said. “And I just see AGI kind of in that light of, yeah, it'll help boost some economic numbers, but that's not going to make life necessarily meaningful for people. And maybe we'll more strongly reject these principles that life is all about GDP.”

• And in this, perhaps the philosophical root of AI — at least in its hyped potential to replace human workers — is one of the oldest philosophies explored by man: what is the meaning of life? 

This journey to find meaning has played out in science fiction as well as our own history of industrialization, both of which point to a theme, according to Giansiracusa, of “humanity rising above the machine.” 

A poll before you go

Thanks for reading today’s edition of The Deep View!

We’ll see you in the next one.

Your view on Google’s Olympic AI ad:

40% of you said it could’ve been great, but it missed the mark. 19% of you said it’s a terrible idea and 15% of you said you hate everything about it.

I hate everything about this:

• “Teach kids not to ever think, let AI think for them.”

Could’ve been great:

• “The receiver should be able to discern the writing of an eight-year-old and expect nothing less.”