- Verses Over Variables
- Posts
- Verses Over Variables
Verses Over Variables
Your guide to the most intriguing developments in AI

Welcome to Verses Over Variables, a newsletter exploring the world of artificial intelligence (AI) and its influence on our society, culture, and perception of reality.
AI Hype Cycle
Decoding the Matrix: Anthropic’s New Economic Index and What it Really Means for Your Job
Ever find yourself doomscrolling about AI taking over your job? We get it. The line between sci-fi and reality is blurring, and instead of killer robots, we've got really good coding assistants. Anthropic, the folks behind the Claude.ai, decided to do something radical. Instead of asking people if they're using AI, they looked at how they're using it. The result is the Anthropic Economic Index, a snapshot of AI's current (and evolving) impact on the workforce.
These were our main takeaways:
Software Devs and Writers, Unite! (Under the AI Banner): Right now, the biggest users of AI are in software development and technical writing. Software development represents a whopping 37.2% of Claude's usage. Think code debugging, software modification – basically, all the stuff that makes your apps work (and occasionally glitch). Arts and media follow next at 10.3%.
Augmentation Nation: AI isn't just stealing jobs (yet). In fact, it's playing nice with humans, for now. The Index found that AI is used more for augmentation (57%), helping humans do their jobs better, than for pure automation (43%).
The Middle-Class Squeeze (But Not How You Think): Interestingly, AI usage is highest in mid-to-high wage jobs, like, you guessed it, programmers and copywriters. Both the super low-paying and super high-paying gigs see less AI action. Maybe because AI can't shampoo hair or deliver babies (yet).
Anthropic is refreshingly upfront about the study’s limitations. They acknowledge that they can’t distinguish between someone using Claude for work or their next great novel. Their data only covers Claude.ai's Free and Pro plans, not their enterprise level API users. And yes, Claude’s reputation for coding excellence might skew the numbers toward software development.
This research confirms what we suspected: AI is already making its presence felt in the workplace. But it also highlights the unevenness of that impact. Some professions are riding the wave of AI augmentation, while others remain largely untouched. The real question isn't whether AI will change work – it's how we'll adapt to work alongside it. This shift isn't about displacement; it's about evolution. As we move forward, the key won't be avoiding AI but understanding how to harness it effectively.
The AI Paradox: The More We Automate, the More We Value Humans
We live in the age of AI worship. ChatGPT writes our emails, Midjourney creates our art, and algorithms shape our daily lives. AI has become our digital companion, promising to solve every problem and automate every task. But something unexpected is happening: the more we automate, the more we crave human connection. This isn't just post-pandemic nostalgia talking. After years of Zoom calls and digital interactions, we're witnessing a fundamental shift in how we value human presence. The very technology that promised to make human involvement obsolete is actually highlighting what makes us irreplaceable.
The business world is catching on. Take Klarna, the fintech giant. After replacing much of their customer service with AI, CEO Sebastian Siemiatkowski made a striking reversal. "In a world of AI," he tweeted, "nothing will be as valuable as humans!" The company is now reimagining its entire approach to prioritize human interaction – a telling sign of where the market is heading.
What we're seeing is a redefinition of luxury. In a world where AI-generated content becomes commonplace, human-crafted experiences stand out. The handwritten note. The personally curated selection. The thoughtfully designed space. The conversation that goes off-script in the most delightful way. These human touches aren't inefficiencies to be eliminated – they're differentiators to be celebrated.
This shift has profound implications for both businesses and individuals. Companies that understand this paradox are already adapting, creating hybrid experiences that combine AI efficiency with meaningful human interaction. They're investing in emotional intelligence training, relationship-building skills, and the art of personal service. Meanwhile, individuals are rediscovering the value of distinctly human skills: creativity, empathy, complex problem-solving, and the ability to build genuine connections. The future of work isn't about competing with AI – it's about complementing it. While AI handles the repetitive and computational tasks, humans can focus on what we do best: understanding context, reading between the lines, providing emotional support, and finding creative solutions to complex problems. This isn't about choosing between efficiency and humanity; it's about using technology to enhance our human capabilities rather than replace them.
Here's the beautiful paradox: AI's greatest contribution might be showing us what makes us irreplaceably human. As automation becomes ubiquitous, human creativity, empathy, and connection will become our most precious commodities. In this automated world, the ultimate luxury won't be more technology – it will be the authentic human touch.
Back to Basics
Holmes vs Watson: Choosing Between OpenAI’s GPTs and Reasoning Models
To say the AI world is constantly evolving is an understatement. It's more like a hyper-accelerated explosion of new models, techniques, and capabilities. Keeping up is a challenge, to put it mildly. And one of the biggest hurdles is figuring out which AI model to use, when, and why. It's no longer about finding the single "best" model; it's about understanding a diverse ecosystem and choosing the right tool for each specific task. OpenAI has recently clarified this conundrum by explaining the fundamental differences between their GPT series and their newer reasoning models (the "o-series").
Think of GPT models as highly skilled workers – they're excellent at executing tasks, following instructions, and producing content. They're the workhorses of the AI world, powering everything from chatbots to code generation. But when it comes to complex planning or strategic thinking, they sometimes fall short. Enter the reasoning models (the o-series). These are the strategists, the problem-solvers who excel at breaking down complex challenges and navigating ambiguous situations. If GPTs are the capable workers, reasoning models are the architects and planners.
The core distinction is this: GPT models are amazing at doing. They execute tasks, follow instructions (mostly), and churn out text (or images, or code) with impressive speed and, increasingly, affordability. The o-series models, on the other hand, are all about planning. They're the strategists, the architects of solutions, the ones who can navigate ambiguity and make decisions based on incomplete or messy data.
Think of it like building a house. You could have GPT handle every single step. It might even do a decent job, especially if you provide very detailed blueprints. But if you want a truly well-designed, resilient structure, you'd probably bring in an architect (the o-series model) to plan things out before you hand the actual construction over to the builders (GPT).
GPT Models Excel At:
Generating content (text, code, or creative writing)
Following clear instructions
Executing well-defined tasks
Processing and responding to straightforward requests
Handling routine interactions and communications
Reasoning Models Shine When:
Dealing with ambiguous problems
Searching through large datasets for specific information
Planning complex solutions
Making strategic decisions
Handling tasks that require deep analytical thinking
The real magic, as OpenAI themselves hint, lies in combining these different model types. Imagine a workflow where an o-series model analyzes a complex legal contract, identifies the key clauses and potential risks, and then delegates the task of summarizing those risks in plain English to a GPT model. Or picture an AI assistant that uses o1 to plan your entire week, taking into account your scattered emails, calendar appointments, and vaguely worded to-do lists, and then uses GPT-4o to actually draft those emails and schedule those meetings.
The industry is moving toward more integrated approaches. Anthropic is reportedly developing hybrid models that combine GPT and reasoning capabilities, while Gemini already offers both in a single package. Some developers are even creating tools that let you mix and match different models' strengths – like pairing DeepSeek's reasoning capabilities with various GPT models.
This isn't just about choosing the "right" model for a specific task. It's about building systems that leverage the strengths of both types of intelligence, creating a kind of AI synergy that goes beyond what either could achieve alone.
Tools for Thought
OpenAI’s 4o Gets a New Vibe
OpenAI has been tinkering under the hood again, and this week's update to the GPT-4o model is all about personality and a commitment to open inquiry. As Andrej Karpathy put it, the new GPT-4o feels "a lot more chill / conversational," like chatting with a friend, not your HR department. Expect a bit of sass, better emotional responses ("That's frustrating!" is now in its vocabulary), and, yes, perhaps an overabundance of emojis. Sam Altman, OpenAI's CEO, even boasted that GPT-4o is now the "best search product on the web"—a statement that included a not-so-subtle jab at the head of Perplexity. But this update goes beyond surface-level charm. OpenAI also significantly revised its Model Spec, explicitly embracing intellectual freedom. This is crucial because it means the AI is designed to empower users to explore, debate, and create without arbitrary restrictions, regardless of how controversial or challenging the topic. The updated spec encourages the model to engage with any perspective, and it's been released under a Creative Commons license – meaning it's completely in the public domain, encouraging others to build upon it. Altman also outlined OpenAI's roadmap, promising to make AI "just work" for users, starting with the elimination of the much-maligned model picker. The near-term plan includes GPT-4.5 ("Orion") as the last non-chain-of-thought model, followed by a unified intelligence system that merges the best of their current models. The ultimate goal is GPT-5 (integrating "o3" technologies like voice, canvas, search, and deep research), though a precise timeline remains under wraps. It appears that OpenAI is making strides to be more open and transparent.
Intriguing Stories
Forget the Bros: These Women are Building the AI We Need
If you are like us, it is getting hard to keep up with all of the AI advancements. So today, we wanted to spotlight two exceptional women, Murati and Raiza Martin, who are not just riding the AI wave – they're shaping it.
Mira Murati: Beyond OpenAI - Building Aligned, Accessible AI: As the former CTO of OpenAI, Murati was instrumental in bringing us ChatGPT and DALL-E. Now, she's heading up Thinking Machines Lab, with a mission to make AI not just powerful, but customizable, understandable, and accessible. Murati's the CEO, and she's assembled an all-star team including OpenAI co-founder John Schulman as Chief Scientist and Barret Zoph as CTO. Thinking Machines Lab is focused on building multimodal AI that collaborates with humans, across fields from science to programming. They want to democratize AI, moving it beyond the closed doors of a few major research labs. What sets Thinking Machines Lab apart is its structure and approach. Research and product teams will work hand-in-hand. They're not just building cool tech in a vacuum; they're focused on real-world applications, from the ground up. Unlike companies hyper-focused on niche areas, Thinking Machines Lab is aiming for broad applicability – AI that adapts to human expertise, not the other way around.
Raiza Martin: From NotebookLM to Talking to Your Business: While leading Google's NotebookLM team, Raiza Martin helped revolutionize how we interact with documents and information. Now, she's taking on an even bigger challenge: teaching AI to understand the complex language of business. Martin's latest venture, alongside designer Jason Spielman and engineer Stephen Hughes, tackles a fundamental question: What if you could have a conversation with your business? Not just its data, but its entire ecosystem of customer interactions, market trends, and operational insights. Her approach is refreshingly practical. Instead of chasing abstract AI capabilities, Martin focuses on solving real-world business problems. The prototype she recently previewed on Twitter demonstrates this philosophy in action – it's a system that seamlessly connects disparate business data sources to extract meaningful customer insights in minutes, not days. What sets Martin's work apart is her understanding that business intelligence isn't just about collecting data – it's about making that data accessible and actionable. Her system acts as a translator, converting complex business metrics and customer interactions into clear, actionable insights that anyone in an organization can understand and use. While Martin hasn't revealed all the details of her new venture, her track record at Google and her current direction suggest something groundbreaking. She's not just building another business tool – she's reimagining how businesses understand and act on their own data.
Mira Murati and Raiza Martin represent a vital shift in the AI landscape. They're pushing beyond the headlines of giant models and theoretical breakthroughs. Murati is focused on the fundamental nature of AI, ensuring it's aligned with human values and widely accessible. Martin is tackling the everyday challenges of integrating AI into our work lives, making it a seamless and intuitive partner. We're eagerly watching both of these pioneers, as they shape a future where AI is not just powerful, but a force for collaboration and progress.
Art, Algorithms and Angst?
Christie's, the auction house that's seen everything from Renaissance masterpieces to, well, a $432,000 AI-generated portrait that looks like a blurry Victorian gentleman, is hosting "Augmented Intelligence," their first-ever auction dedicated entirely to art created with the help of Artificial Intelligence. But, before you picture robot arms wielding paintbrushes, the reality is, predictably, a lot more complex – and a whole lot more controversial. The auction, which runs from February 20th to March 5th, features digital art, sculptures, paintings, and prints. Christie's is positioning this as a celebration of AI as a collaborative tool, an artistic sidekick that enhances human creativity rather than replacing it. Think of it as a digital muse, whispering possibilities into the artist's ear (or, you know, processing algorithm). But (and it's a big but), not everyone is buying the "kumbaya" narrative. This auction has landed smack-dab in the middle of a fiery debate about the ethics of AI-generated art. An open letter, signed by nearly 6,000 people, demanded the auction's cancellation because the AI models used to create some of these artworks were allegedly trained on copyrighted material without the original artists' permission. Christie's, for their part, is trying to navigate this PR storm with a carefully crafted message. They emphasize that the featured artists have "robust multidisciplinary practices" and are using AI to enhance, not replace, their human creativity. It's important to remember that Christie's isn't new to the AI art game. Back in 2018, they sold the aforementioned "Portrait of Edmond de Belamy," a generative adversarial network (GAN) creation that sent shockwaves through the art world. That sale was a watershed moment, signaling that AI-generated art had arrived on the big stage. This new "Augmented Intelligence" auction, then, isn't just a one-off event. It's a continuation of a trend, a sign that AI is becoming increasingly intertwined with the art world, whether we like it or not.
Forget Spreadsheets, AI is Painting Now
We've all heard the hype about AI transforming industries, boosting quarterly earnings, and generally taking over the world (probably while wearing a perfectly tailored suit). AI isn’t just revolutionizing big business, it is impacting everyday people. Take, for example, the local nail technician down the street in OpenAI’s Ten Tiny Canvases.
OpenAI recently showcased how nail artist Tabytha "Taby" Scott is using ChatGPT to supercharge her creative process. We are not talking about automating manicures (yet!). We're talking about an AI brainstorming buddy. And honestly, we loved this video – it perfectly reflects how many of us are already weaving AI into our daily lives. We use AI in a similar way, in fact, we are tapping into advanced voice mode constantly. Taby, a licensed nail tech with a lifelong passion for art, found herself facing a creative block. Instead of staring blankly at a wall of polish, she turned to the ChatGPT app on her phone. By describing a client's general idea, Taby received instant, specific suggestions. Taby's story highlights the accessibility of AI tools for everyday creativity. She's not a coder. She's an artist. And ChatGPT is acting as her digital muse, helping her translate vague concepts into ten tiny, dazzling canvases. ChatGPT helps Taby narrow down an infinite number of design possibilities to a manageable amount. Then Taby can add her own expertise to make a truly unique piece of art. The lesson here is that AI tools shouldn’t be intimidating, they are becoming more intuitive and helpful for specialized tasks. Jump in, and ask a friend!
— Lauren Eve Cantor
thanks for reading!
if someone sent this to you or you haven’t done so yet, please sign up so you never miss an issue.
if you’d like to chat further about opportunities or interest in AI, please feel free to reply.
if you have any feedback or want to engage with any of the topics discussed in Verses Over Variables, please feel free to reply to this email.
banner images created with Midjourney.