- The Deep View
- Posts
- ⚙️ When the glitter of an economic boom clashes with the environment
⚙️ When the glitter of an economic boom clashes with the environment

Good morning. OpenAI is remarkably hungry for more data. Beyond rumors that it’s working on developing a social media app — really, little more than a relentless data source — the startup said that it would be interested in buying Chrome from Google if a court orders a spin-off.
Data, data, data, data.
— Ian Krietzberg, Editor-in-Chief, The Deep View
In today’s newsletter:
🧠 AI for Good: Detecting Alzheimer’s
📊 Tesla did even worse than analysts were expecting
🌎 When the glitter of an economic boom clashes with the environment
🎙️ Podcast: Programmable plants
For the latest episode of The Deep View: Conversations, I sat down with Brad Zamft, the co-founder and CEO of Heritable Agriculture, a startup that’s leveraging machine learning technology to program plants.
This is a special one, to me — I don’t often get an in to talk about regenerative agriculture and the importance of sustainable farming practices (but when I’m not talking about AI, I’m staring at trees and lamenting the destruction of our planet).
Fitting that this episode came out on Earth Day, though that was really just a happy accident.
Give it a watch/listen.
AI for Good: Detecting Alzheimer’s

Source: Unsplash
Today, there is no cure for Alzheimer’s Disease.
But that doesn’t mean that early diagnosis isn’t necessary. Researchers have found that it might be possible to either delay or prevent the development of dementia in some patients through specific interventions, including drugs that slow the progression of the disease, drugs that treat memory loss symptoms and non-drug therapies, like making art.
What happened: On a mission to improve early Alzheimer’s diagnoses, researchers at UC San Francisco last year built a machine learning model that can predict the disease up to seven years before symptoms begin to surface.
The team trained random forest models on the clinical data of more than five million Alzheimer’s patients from UCSF’s Memory and Aging Center. They were looking for patterns that appeared consistently across those patients, paying significant attention to signs that were present in patients with Alzheimer’s, but were not present in patients without the disease. (Random forests are a popular type of machine learning model. Read more about them here).
Thus equipped, the model was capable of predicting Alzheimer’s up to seven years before symptoms set in with 72% predictive power.
Why it matters: The researchers were able to examine predictive factors for both men and women separately, which, considering the increased rate at which women tend to get Alzheimer’s, represents an important discovery. They found that hypertension, high cholesterol and vitamin D deficiencies act as predictive factors for both men and women, adding that osteoporosis is a predictive factor for women.
“It is the combination of diseases that allows our model to predict Alzheimer's onset,” Alice Tang, the study’s lead author, said in a statement. “Our finding that osteoporosis is one predictive factor for females highlights the biological interplay between bone health and dementia risk.”

Big ideas. Bold strategies. AI-powered growth.
Experience it all live at ZoomInfo’s GTM 25 on May 13th at the Guggenheim Museum in New York.
This exclusive experience brings together go-to-market leaders to explore innovative GTM strategies with industry experts, including ZoomInfo Founder & CEO Henry Schuck and McKinsey & Co. Partner Todd Horst.
Connect with peers, dive deep in interactive breakout sessions, and be the first to see ZoomInfo’s groundbreaking platform innovations.
Tesla did even worse than analysts were expecting

Source: Unsplash
The analysts weren’t expecting much from Tesla this quarter.
Earlier in the month, the company reported a 13% decline in vehicle deliveries, a dip that comes amid massive declines in Elon Musk’s popularity, primarily due to his government-slashing work with DOGE.
Here’s how they did: Tesla reported earnings of only 27 cents per share, below expectations of 39 cents. The company reported revenue of $19.3 billion, below expectations of $21.11 billion (that’s a 9% year-over-year dip).
Tesla declined to provide a growth forecast, saying that it will “revisit our 2025 guidance in our Q2 update.”
Tesla said in its slide deck that “AI is a major pillar of growth for Tesla and the broader economy and key to our pursuit of sustainable abundance.” It sought to affirm investors that the company remains “on track” for a pilot launch of its robotaxi in Austin “by June.”
Among its many headwinds, Tesla warned that “uncertainty in the automotive and energy markets continues to increase as rapidly evolving trade policy adversely impacts the global supply chain and cost structure of Tesla and our peers.”
And though the stock is down nearly 40% for the year, it rose 4.5% during the day, and spiked another 4.5% in after-hours trading, mostly, it seems, on news that Elon Musk will be at least partially returning to Tesla.
He said on the call that, starting next month, his “time allocation to DOGE will drop significantly,” adding that he’ll continue to spend one to two days per week “on government matters for as long as the president would like me to do so.”
Musk added that “there will probably be some unexpected bumps this year,” though he remains “extremely optimistic about the future of the company,” a future that is “fundamentally based on large-scale autonomous cars and large-scale … humanoid robots.”
He said that the financial impacts of autonomy will become “material” sometime next year, adding that Tesla will be producing one million units per year in less than five years.
Musk has been vowing that autonomy is right around the corner for about a decade. It has yet to materialize. I am sure that, since Texas doesn’t have any laws regarding robotaxis, autonomous Teslas of some form or another will launch in the state this year. But that’s no indication of a wider rollout, especially considering the fact that Tesla’s software — despite its name — requires the constant attention of its human driver.
It’s a hurdle that requires redundancy layers, like the radar and lidar that Musk and Tesla are so vehemently opposed to.
Tesla is already facing more than a dozen FSD-related lawsuits.
I don’t expect the rollout to go well.


Forcing its way into education: According to the Washington Post, the Trump Administration is considering implementing an executive order that would integrate generative AI into K-12 schools. Federal agencies would “be asked to partner with the private sector to develop relevant programs in schools.” Subtle.
Another one bites the dust: The Washington Post on Tuesday joined brands including News Corp, Vox Media, Hearst and Conde Nast in inking a content licensing deal with OpenAI. Others of its contemporaries have opted to sue OpenAI instead. The Post is owned by Jeff Bezos.

Google is paying Samsung an ‘enormous sum’ to preinstall Gemini (The Verge).
Google used search data to train its AI models (The Information).
Columbia student suspended over interview cheating tool raises $5.3M to ‘cheat on everything’ (TechCrunch).
Trump’s tariffs will ‘significantly slow global growth,’ IMF warns (Semafor).
4Chan is dead. Its toxic legacy is everywhere (Wired).
When the glitter of an economic boom clashes with the environment

Source: Unsplash
Generally, when we talk about AI and the environment, we also talk about economics.
The throughline is simple. The AI-related problems that plague the environment are the same ones that make the technology so expensive in the first place: masses of (wildly expensive) energy-hungry chips are significantly boosting the energy (and water) consumption of data centers around the world, threatening local ecosystems with water scarcity, air pollution and unreliable electrical grids.
In a report published Tuesday, the International Monetary Fund (IMF) expects the data center power demand of the U.S., Europe and China to grow at an annual rate of 22%, 13% and 10%, respectively, between now and 2030.
The details: In a projection where renewable energy doesn’t scale up fast enough, and where energy infrastructure investment doesn’t get large enough, the IMF said that, by the end of the decade, energy prices could increase by 8%, U.S. carbon emissions could increase by 5.5% and global carbon emissions could increase by 1.2%.
The IMF’s modeling found that AI-related electricity consumption could reach 1,500 terrawatt hours (TWh) by 2030, roughly comparable with India’s total electricity consumption (India is the third-largest electricity consumer in the world).
In 2023, data centers and AI consumed around 500 TWh of electricity, more than double the sector’s consumption in 2015, which was flat until 2019. So, between 2019 and 2023, consumption doubled. And between now and 2030, it could triple.
But, according to the report, all that consumption will be worth it: “the growth in GDP level from AI expansion greatly exceed(s) the fiscal costs,” the report reads. There isn’t much clarity here regarding the legitimate ways in which AI — a broad, fuzzy term — will contribute to GDP growth, beyond the “AI-induced expansion of the IT sector.”
That boost in GDP, however, comes at the cost of higher carbon emissions.
Specifically, the “AI-driven rise in electricity demand could add 1.7 Gt in global greenhouse gas emissions between 2025 and 2030, similar to Italy’s energy-related GHG emissions over a 5-year period.”
Here, Gt refers to gigatons. One gigaton is equal to one billion tons. The average car weighs around two tons (4,000 pounds), so imagine 500 million cars (500 million is more people than live in the U.S.).
This isn’t too much of a concern for the IMF.
“The social cost of these additional emissions represents only a very small portion of the anticipated aggregate economic benefits from AI,” according to the report. “However, they would nonetheless contribute to an already concerning accumulation of emissions. In addition, while the additional emissions will have global impacts, the benefits of AI will likely be unequal both across countries and among different groups within societies, potentially exacerbating existing inequalities.”
Drawing on a variety of different estimates, the report affixes a pricetag to that “social cost,” ranging from $50.7 billion to $66.3 billion, which would only be 1.3% to 1.7% of the “AI-driven increase in real-world GDP between 2025-2030.”

In 2017, just eight years ago, the Earth’s atmosphere broke what the Yale School of Environment called a “startling” record: it exceeded 400 parts per million of carbon dioxide.
The last time this planet had levels that high was around four million years ago, a time when North America was still connected to South America and humans didn’t exist.
Yesterday, on Earth Day, the concentration of CO2 in the atmosphere surpassed 430 parts per million.
I’m not sure how the IMF is calculating its social cost — presumably, they’re quantifying costs related to strain on local health systems due to impacts from air pollution. It’s not clear if they’re including broader costs related to climate change, such as storm damage, flood damage, fire damage and global supply chain interruption and destruction, to name a few.
A study from last year tried to affix a pricetag to all of that, saying that by 2049, climate change will cost the world roughly $38 trillion per year, every year (38 trillion is like 130 Elon Musks).
I have said it before, and I will say it again: AI does not exist in a vacuum. GDP gains or economic disruptions relevant to AI do not exist in a vacuum.
We are at a stage where anything that increases carbon emissions — or prevents us from decreasing carbon emissions — poses a potentially significant threat.
We must not lose sight of that in the face of vague hypotheticals.
The International Energy Agency’s recent report on the future of energy and AI directed plenty of attention to vague, hypothetical ways in which the tech could reduce emissions, but added the vital caveat that there is “currently no existing momentum of Al adoption that would unlock these emissions reductions …”
It’s easy to say that the benefits, either economically or environmentally, will outweigh the costs. But that impression isn’t supported by any clear, rigorous evidence.
The impacts of climate change are happening now, in ways ranging from longer wildfire seasons to longer allergy seasons and worsening storms.
They will get worse.
We, as a society, have to think much harder about the kind of world we want, and how we want to get there. We have to think about what we’re okay sacrificing, and for what, and for whom, and why.
I am concerned with one existential threat.
It has little to do with AI.


Which image is real? |



🤔 Your thought process:
Selected Image 2 (Left):
“Image one subject and composition is cookie-cutter nice.”
Selected Image 1 (Right):
“The words on the piano threw me off.”
💭 A poll before you go
Thanks for reading today’s edition of The Deep View!
We’ll see you in the next one.
P.S. Enjoyed reading? Take The Deep View with you on the go! We’ve got exclusive, in-depth interviews for you on The Deep View: Conversations podcast every Tuesday morning. Subscribe here!
If you want to get in front of an audience of 450,000+ developers, business leaders and tech enthusiasts, get in touch with us here.