hello!
Cognitive Warmup. For lack of a more appropriate descriptor, I still call Canva a unique visual communications platform, and one that I find relevant almost every other day as part of my work.
Their AI layer remains very relevant, which creative os And the Affinity app caters well to different creative workflows, and something co-founder and chief operating officer Cliff Obrecht also illustrated In detail with examples of tools like Canva Code. Now, Canva has done something that could become popular. Or not. Only time will tell. But they add a new language setting called English (chronologically online), and it has become Canva’s 101st supported language.
“Users are increasingly motivating Canva AI with phrases like ‘key character energy,’ and Canva is meeting them halfway, recognizing that the way people speak online can be a functional, familiar shortcut that leads to creativity,” is how Canva explains it. Now, cool kids can use phrases like “slay” and “spill the tea” in their signs and conversations with Canva AI. Worryingly, Canva also reminds us that “By 2030, Gen Z will make up approximately 30% of the US workforce” – and I’m assuming it will be a similar figure in most other countries, including India.
Anyway, I digress. The platform now has 260 million monthly users, and Apple recently announced a competitor Creator Studio (Adobe would be concerned too, I’m sure).
algorithm
There are no simple assessments in the world of AI. This week:
- Decoding AI Browsers and Are They Really Evolved
- Chronicling the Windows 11 update that keeps breaking things
- OpenAI is temporarily testing a unique revenue stream
Where are the AI browsers?
This is more of a philosophical question, given that several months have passed since Perplexity and then OpenAI Issued (Amid much hype and excitement, let me make this very clear) what they call an “AI browser”, which apparently has the ability to perform certain tasks on its own. “A thoughtful partner and assistant,” says Arvind Srinivas, CEO of Perplexity. has described Comet.
This month, OpenAI announced an update to Atlas that adds tab groups and quick suggestions to the Ask ChatGPT sidebar. It was a good reminder that I’ve heard very little about anything significant on this front over the past few months. This, after a mind-numbing campaign that pretty much told us we were using the web browser wrong all this time.
Along the way I also looked at some numbers so you have some perspective to work from.
The latest data from research firm StatCounter still pegs Chrome as the undisputed leader with 71.23% market share, with Apple Safari (14.84%), Microsoft Edge (4.6%; boosted by the fact that it’s preloaded on millions of Windows PCs), Firefox (2.25%), Samsung Internet (1.89%; reflecting Samsung’s smartphone dominance) and Opera (1.83%) topping the list, before it Fall into the “other” category.
Despite all the claims of being the smartest assistant ever, this style is getting very little popularity. This is a fact.
History of broken promises
You might have expected me (and I considered it) to write about the never-ending saga of messy Windows 11 updates in my other weekly newsletter, Wired Wisdom, as I did last week. Pointing towards a downward spiral. This week, I’m moving this conversation here, because…well…Microsoft prefers to call it simply “Agent OS”..
Microsoft has released the January Software Update for Windows 11, which is well complimented by two emergency patches:
- The first to restore a critical functionality to millions of Windows PCs, which will let them close properly again.
- The second fixes what the previous update inadvertently introduced – a feature in which “some applications may become unresponsive or experience unexpected errors when opening or saving files in cloud-backed storage such as OneDrive or Dropbox”.
Throughout my 18 years and counting in this profession, I have felt time and again that Microsoft is not a company that is very receptive to feedback, especially from people who don’t necessarily think that Windows OS – and now AI Copilot – are the greatest things to ever grace God’s green earth. If there aren’t people they don’t necessarily like, at least they can listen. One of their biggest PC partners.
I find the words of Reed Southen, known for his visual effects work on The Hunger Games (2012), Blue Beetle (2023) and Transformers: The Last Knight (2017) quite apt for this situation – “Don’t they test anything anymore? What a bunch of vibe coding clowns.” I assume Southen’s frustration comes from one of his critical workflow machines being stuck? But okay, an “agentic OS” would be Windows 11.
Advertising supported, and revenue sharing.
What was once touted by CEO Sam Altman as a “last resort” has become the new business model. After the commercials, there is a light hint on revenue sharing. Or as I have summarized it, and perhaps you will agree, “A Finder’s Fee”.
It’s clear, OpenAI is open to the idea of customers taking part in AI-assisted searches. It is equally clear that the AI company urgently needs more revenue sources. The conversation about ads in ChatGPT points to this.
OpenAI hopes to begin to transform its relationship with enterprise customers from merely a “tool provider” to one of a long-term infrastructure partner and they hope to eventually include AI agents deployed in the workplace. And they would like a cut of that achievement.
Thinking
“I think one of the good choices we made early on was to become a company that focuses on the enterprise rather than the consumer. It’s easier to choose a business model where you have less to fight with your own business incentives.
I am very concerned about consumer AI because of the need to maximize engagement. It leads to a slope. We’ve seen things related to ads from some other players. Anthropic doesn’t work that way or need to work that way. We just sell things to businesses and they are priced directly. We don’t need to monetize a billion free users or maximize engagement just because we’re in a death race with another big player. — Dario Amodei, CEO of Anthropic at the World Economic Forum
I had analyzed in detailInteresting exchange between Amodei and Google DeepMind co-founder and CEO Sir Demis Hassabis, involving radically different views on the approach, AGI timelines, and the pace of AI development.
The conversation we are having here is about what Amodei said at another event in Davos, but it should not be ignored. Think of it this way – Amodei’s outline is not revealing what she criticizes, but rather what she clearly accepts.
Consumer AI, as it is currently being built, is structurally incentivized towards noise. Engagement-first economics rewards quantity over value, velocity over authenticity, and innovation over utility. Once you accept advertising, virality, or ad-supported freebies as the main distribution logic, you also accept that slope is not a bug – it becomes a feature, especially for engagement.
By the way, Microsoft CEO Satya Nadella gets angry if someone calls an AI generation “sloping.” In fact, he even wrote a blog earlier this month, and I quote- “We need to go beyond the arguments of sloppiness versus sophistication and develop a new balance in the context of our “theory of mind” that is responsible for equipping humans with these new cognitive amplifying tools as we relate to each other.”
The ‘microslop’ trend on X should not be pleasant.
Anyway, I digress. Back to the point.
Anthropic emphasizes that selling AI directly to businesses forces a different kind of discipline, but it could clearly be argued that this is more a defensible structure than an ethical benefit. The incentives are clear, and the path to development is well-lit. There is a uniqueness in all this. In that sense, Amodei is right – it’s easier to design “responsible AI” when your business model doesn’t depend on keeping a billion users scrolling, engaged, and eventually paying for a subscription (or advertising on their interaction window).
A reality check: If consumer AI, as Amodei points out and I agree with him, is inherently prone to degradation, then we could be sleepwalking into a two-tier AI future. The other side of this coin is enterprise AI, where seriousness prevails. Therefore, businesses and enterprises will get relatively more focused and accurate tools with better guardrails and perhaps accountability. Consumers, on the other hand, get dopamine loops, synthetic-induced clutter, and a bleak outlook on life.
Amodei clearly positions Anthropic as breaking out of the “death race,” and that direction is something that has always impressed me about Anthropic (even before they started explicitly stating their different approach).
The question for an AI race that still wants to appeal to consumers is not whether engagement-driven AI yields sloppiness – that already happens, regardless of what Mr Nadella might want to say about it – but whether the industry is willing to accept some realities that are fundamentally incompatible with their vision of revenue-generating intelligence. If the answer is no, enterprise-first AI could be a safe haven.
Produced by Tushar Deep Singh.
Neural Dispatch is your weekly guide to the rapidly evolving landscape of artificial intelligence. Each edition provides curated insights on the critical technologies, practical applications and strategic implications shaping our digital future.
Want this newsletter sent to your inbox? Subscribe here.
