- Everyday AI
- Posts
- Ep 667: Aligning AI With Climate And Business Goals
Ep 667: Aligning AI With Climate And Business Goals
Meta makes big AI acquisition, OpenAI could release GPT-5.2 next week, Google upgrades Deep Think and more.
Everyday AI
December 05, 2025
Outsmart The Future
Today in Everyday AI
8 minute read
🎙 Daily Podcast Episode: AI’s impact on the climate is often misunderstood. We break down how companies can align AI adoption with ESG and climate goals without sacrificing business performance. Find out more in today’s show and give it a watch/listen.
🕵️‍♂️ Fresh Finds: Google’s big AI agent upgrades, why Foxconn grew 26% in one month, NotebookLM’s small (but major) upgrade. Read on for Fresh Finds.
đź—ž Byte Sized Daily AI News: Meta makes big AI acquisition, OpenAI could release GPT-5.2 next week, Google upgrades Deep Think and more. Read on for Byte Sized News.
🧠Learn & Leveraging AI: AI isn’t killing your climate goals. Your inefficient operations are. We break down why the “AI is bad for the planet” narrative misses the real problem. Keep reading for that!
↩️ Don’t miss out: Anthropic and Snowflake $200M AI Deal, Google makes splash with Gemini Workspace Studio, OpenAI makes big nonprofit investment and more Check it here!
The Future of AI Agents: Will there be more Agents than humans?
How can you scale AI at the enterprise, yet still hit your climate goals?
And can heavy AI usage and an enterprise's ESG mission co-exist?
Ashutosh Ahuja lays it out for us.
Aligning AI With Climate And Business Goals -- An Everyday AI Chat with Jordan Wilson and Ashutosh Ahuja
Also on the pod today:
• AI’s water usage myth 💧
• Predictive waste sorting with AI 🗑️
• Nest thermostat: AI energy saver 🌡️
It’ll be worth your 27 minutes:
Listen on our site:
Click to listen
Subscribe and listen on your favorite podcast platform
Listen on:
Here’s our favorite AI finds from across the web:
New AI Tool Spotlight – Documentation.AI creates documentation automagically, Orca is an AI agent for game dev and Index for ChatGPT gives long chats an index.
Government and AI — Even the FDA is getting in on Agentic AI.
AI Competition — AI "godfather" predicts Google leapfrogs OpenAI with homegrown chips.
AI Bubble — When and if the AI bubble bursts, who will still be around?
AI and the Economy — Foxconn’s revenue jumped 26% in November, fueled by booming demand for AI server racks. See how.
NotebookLM Updates — NotebookLM is expanding custom instructions.
Google AI Limits — For paid subscribers, Google is upping limits for its new Antigravity agent platform.
Google Agent Upgrades — Google’s Opal engine is about to get way smarter, adding direct Gmail, Calendar, and Drive automation plus a new Agent for multi-step problem solving.
AI Progress — AI critics say progress is stalling, but the data shows explosive growth and real-world impact. Are we in denial about how fast super-intelligent machines are changing our lives?
1. Meta Snaps Up AI Wearables Startup Limitless đź‘“
Meta is making headlines today with its acquisition of Limitless, the startup behind a conversation-recording wearable pendant, signaling a fresh push into AI-powered consumer hardware.
The move follows Meta's recent hiring of ex-Apple design chief Alan Dye, showing the social giant is serious about next-gen devices. Limitless’ technology will fuel Meta’s ambitions for smarter AI wearables, though new device sales are now paused while current users face updated privacy terms. With more than $33 million already invested in Limitless, Meta is clearly betting big on the future of AI-driven productivity tools.
2. Report: OpenAI Races to Release GPT-5.2 in "Code Red" AI Showdown ⚔️
OpenAI is reportedly moving up its GPT-5.2 release to December 9th, after CEO Sam Altman called a "code red" to respond to fierce competition from Google’s Gemini 3 and Anthropic. Sources say GPT-5.2 aims to close the gap on Google's new model, which recently dominated AI leaderboards and caught the attention of tech heavyweights.
If plans hold, the update could debut as early as next week, but OpenAI is known for shifting schedules based on internal hurdles and rival announcements. Expect ChatGPT improvements to follow, with a new focus on speed, reliability, and customization as the AI arms race heats up.
3. NYT sues Perplexity in fresh shot at AI over copyright đź‘®
The New York Times has filed a new lawsuit in federal court accusing AI startup Perplexity of illegally scraping its articles, videos and podcasts and sometimes spitting out near-identical copies as answers. The Times says it supports AI but not using its journalism without a license, while Perplexity is casting the dispute as part of a long tradition of media suing new tech.
Coming on top of the Times’ separate case against Microsoft and OpenAI and Anthropic’s $1.5 billion copyright settlement with authors, this fight highlights how copyright is quickly becoming one of the central legal flashpoints of the AI boom.
4. Meta Strikes Fresh AI Deals with Major News Outlets đź“°
Meta just inked new multiyear data agreements with big-name publishers like CNN, USA Today, Fox News, and Le Monde to ramp up real-time news answers in its AI chatbot, Axios reports. After stepping back from news payouts, Meta is now paying publishers again to fuel its AI with verified content across Facebook, Instagram, WhatsApp, and Messenger.
The move signals a sharp pivot: instead of being a news-sharing hub, Meta’s platforms are shifting focus to instant information and viral videos. By expanding these deals, Meta is betting that real-time, trustworthy news will keep users engaged as it races to lead the AI chatbot game.
5. Anthropic rolls out AI-powered mass interviewer 🗣️
Anthropic has just launched "Anthropic Interviewer," an AI tool that can run thousands of in-depth interviews at scale to study how people actually feel about and use AI at work, and it is already live for many Claude.ai users this week via a pop-up invitation.
In an initial study of 1,250 professionals, a new report from Anthropic finds that workers, creatives, and scientists are broadly optimistic about AI’s productivity boost but remain anxious about job identity, creative displacement, and trust in AI’s reliability.
6. Google’s new “Deep Think” AI mode hits paying users �*
Google is rolling out its Gemini 3 Deep Think mode to AI Ultra subscribers right now, marking the company’s latest move in the high-end AI arms race. The $250-a-month upgrade in the Gemini app promises more advanced problem solving using what Google calls “parallel reasoning,” and the model reportedly scored 41 percent on the notoriously tough Humanity’s Last Exam benchmark.
The launch comes on the heels of Gemini 2.5 and arrives as demand for Google’s AI tools surges so much that the company briefly capped its Nano Banana image generator for free users.
It caused a total panic.
Companies literally paused their AI rollouts because they thought using LLMs would torch their ESG goals.
But here’s the irony.
The architecture lead at Starbucks just confirmed that the "AI is bad for the climate" narrative is looking at the wrong side of the balance sheet.
You are worried about the cost of the query.
You should be worried about the massive inefficiency of your current operations.
So on today’s show, we unpacked why the energy-guzzling narrative is backward, how "lazy" infrastructure is secretly burning thousands of dollars a week, and why your expensive machinery is begging for predictive maintenance.
If you are skipping AI to be "green," you are losing green.
1. The carbon ledger is upside down 📉
Most leaders look at AI purely as a consumption metric.
They see data centers burning energy and cooling systems chugging water.
Wrong.
Ashutosh Ahuja, Enterprise Architecture Lead at Starbucks, argues that this view misses the massive savings on the other side of the ledger.
Think about the Nest thermostat.
It uses machine learning to learn your habits.
Yes, that compute costs energy.
But the resulting drop in your HVAC usage saves infinitely more energy than the model ever consumed.
It’s a net positive for the planet.
We see this same pattern with computer vision in waste management.
Ashutosh described a project where students used AI to sort trash from recyclables automatically.
That one camera system reduces the diesel fuel needed to transport rejected loads and cuts down on massive industrial sorting energy.
So stop looking at just the cost of the query.
You need to look at the efficiency of the output.
Try This:
Instead of letting someone quote that “AI is cooking the planet” meme, send them a fresh stat: KPMG’s November 2025 global survey of 1,200 execs across 20 markets found that 87% say AI is central to achieving net-zero goals, and 96% believe clean energy can meet AI’s power demands if prioritized.
2. Lazy infrastructure is burning your cash đź’¸
You might be bleeding money on cloud costs without even knowing it.
Ashutosh shared a story from his time consulting for Cigna where they moved from on-premise servers to AWS.
They turned on auto-scaling.
This means the servers automatically spun down when nobody was using them.
The result?
They saved thousands of dollars per week immediately.
Here is the connection most people miss.
Every dollar you save on cloud compute is also a direct reduction in your carbon footprint.
If you aren't using auto-scaling, you are essentially leaving the lights on in an empty stadium every single night.
Yikes.
It’s bad for the planet.
It’s even worse for your P&L.
Try This:
Log into your cloud provider’s billing dashboard right now—whether that is AWS, Azure, or Google Cloud.
Sort your instances by "utilization rate" over the last 30 days.
Identify any server that is running 24/7 but has usage spikes only during business hours.
Tag those specific instances for an "auto-scaling pilot" and send that list to your dev ops lead.
You could shave 30% off that bill by simply letting the robots turn the lights off for you.
3. Reactive maintenance is a rookie mistake 🛠️
Some industries are still running massive operations based on the "break-fix" model.
They buy a $500,000 machine and wait for it to explode before they touch it.
This is insane.
Ashutosh pointed out that the renewable energy sector specifically is sitting on a goldmine of efficiency data that they are ignoring.
They have half-million-dollar bio-waste processors running until failure.
This is where AI changes the game physically.
By using IoT sensors and simple predictive models, you can service that machine two weeks before it breaks.
This saves the massive carbon cost of manufacturing replacement parts.
It also saves the rush-shipping logistics to get the parts there.
And obviously, it saves you the downtime revenue loss.
If you are waiting for things to break, you are actively choosing to burn money.
Try This:
Walk through your facility (or check your asset list) and identify your single most expensive piece of equipment.
Find the maintenance log and look for the last three times it failed unexpectedly.
Ask your operations manager: "What data point or noise did this machine make right before it died?"
Take that one specific signal—vibration, heat, or sound—and ask a vendor for a simple IoT sensor that tracks just that one metric.
You just moved from reactive chaos to predictive control in one afternoon.
Reply