🌟 Editor's Note. This week…..
 
Perplexity pulls back from ads, ChatGPT secure $110bn of funding in a monumental raise whilst announcing a new $20,000 a month subscription. Google updates Nano-Banana and Anthropic is upsetting the US Government. Jack Dorsey claims that AI is to blame for 4,000 layoffs in his company Block (spoiler - there’s more to this story than it initially appears).
Whilst all this is going on, a survey is showing the volume of people who are keeping quiet about their AI usage at work….

 _{The $20,000 ChatGPT subscription}

OpenAI is launching a $20,000 per month subscription. What do you get for this price? Autonomous agents capable of executing PhD-level work. Whilst the full reveal is of this new subscription is yet be seen, it is likely there will be a number of different tiers ranging from low $000’s for a general knowledge worker up to $20k for research-level agents.

This announcement came days before the most recent funding round of OpenAI, a raise of $110bn, giving the company a $730bn valuation - one of the biggest raises in history for a company at this stage pre going public. How much OpenAI will rely on subscriptions and how much on ads remains to be seen. It’s a pretty niche market - perhaps some hedge funds, tech companies or consulting firms - that would be able to find the value at $20,000 a month, versus what a good employee with Claude’s Opus or OpenAI's Codex could do. If, however, an AI agent can perform the work of multiple analysts or engineers, perhaps this cost becomes justifiable.

Takeaways for Leaders

The subscription model for more powerful AI is very much the theme of the past couple of months. What can be achieved with the paid-for models will be a differentiator for businesses now and in the future. If you're operating with a free version of Gemini or ChatGPT or Co-Pilot, you're at a disadvantage versus those who are accessing more powerful models.

There may be a temptation for some organisations to think of replacing part of their staff with these subscriptions. But the common theme on human augmentation with AI is very relevant here: any subscription is only a multiplier if there are people who can put it in place and make it work within the context of the business.

_{Perplexity Walk Back From Ads, Release ‘Perplexity Computer’}

Perplexity, the AI “answer engine” which combines live web retrieval with LLMs, has moved away from advertising, whilst launching a new service called Computer. This is an agentic product, coordinating across 19 different AI models like a general-purpose digital worker.
In layman's terms, to give you an outcome, it executes multiple steps across several platforms rather than just answering a question. For instance, it would be able to collect stats, financial or legal data, then produce analysis and then output results as finished websites or dashboards or visualisations.

There are no definitive figures on Perplexity's user base and revenue. What's interesting is that at least for now, they are turning their backs on advertising as a source of revenue and going all-in on subscriptions.
Advertising as a source of revenue is often the most obvious one for tech companies to go after. The sheer scale of the digital revenues of just Meta and Google is an example of how powerful the use of data has become in driving the outcomes that advertisers want.

Takeaways for Leaders

AI companies have plenty of data and the ability to use it. And yet we're seeing a subscription play from both Anthropic and OpenAI and now Perplexity. It could be that these companies are experimenting with different approaches as they figure out their paths to profitability. Or potentially they may be having a tough time entering the advertising market and believe that organisations will see value in paying for subscriptions instead.

The success of Claude (subscription only, not offering ads) in recent months may have helped inform this move. For leaders, a combination of different options to operate alongside your teams to augment their work is becoming clear. Also becoming clear is the gap that's emerging between the paid-for and the non-paid-for models, as well as the volume of change in a short period of time between the different models.

Having your teams experiment with the different options and creating an environment where they can learn safely, whilst not locking yourself into one single enterprise contract is the most appropriate path for most companies right now.

For more information click here

_{Is AI Your Team’s Little Secret in the Workplace?}

According to a recent survey from the software and cybersecurity company Ivanti, one in three workers using AI are keeping it a secret from their employer.

The survey also covered a number of other things about workplace preferences and behaviours. Whether people are more productive in the office or at home and how that differs by age is an interesting call-out (spoiler alert: Gen Z and Baby Boomers are both well aligned on how they both feel unproductive outside the office). There is also another interesting chart about behaviours to look busy internally. For instance, 20% of people admit to jiggling their mouse to pretend they are being active, while 7% admit to having faked a phone call to project that they are busy.

Takeaways for Leaders

The survey sample is relatively small compared to some other surveys we've seen recently. It is limited to office workers and people in the software and cyber security area specifically. Still, there are still some call-outs that leaders should be aware of.

If you're not creating safety among your team, AI usage will still go on, but will be ‘underground’. This is both a risk to security but also a massive waste if teams are not sharing their learnings with each other. The volume of employees surveyed that said they don't have an AI usage policy is quite staggering. With AI usage growing rapidly, if it's something that hasn't been implemented in your company already, this should be an immediate priority.

You can find their research in this link here

🔥 In Case You Missed It…

Tech Entrepreneur Jack Dorsey - the ex-CEO of Twitter and co-founder and CEO of Block - has announced 4,000 layoffs, explicitly linking this move to AI.

Industry commentary pushed back that the layoffs look like over-hiring correction and margin discipline with an “AI” narrative layered on top. Effective AI tools are very new and becoming more effective every day, as we’ve seen in the developments in Claude, OpenAI and Perplexity over recent weeks.

Most organisations are still figuring out how to structure teams around AI. It’s hard to believe one company has already determined which roles are no longer needed. And so, whilst the headlines and naysayers will shout that this is more evidence of AI scrapping jobs, it’s likely that there is a little more to this move than just AI.

🏆 Tools, Podcasts, Products Or Toys We’re Playing With This Week

Claude. Again.  

The last newsletter had an update on the different types of Claude we were playing with and we're still playing with it so it deserves a second mention.

Having recently received 8 pages of output from blood tests, I wanted to have an easy way to compare my metrics over time. I put the past sets of medical data into a folder, gave Claude access to the folder, and asked it to build me a picture of my results.

It's worth challenging yourself to use AI to see if you can find little improvements that would help your day-to-day work or day-to-day life outside of work. This build took me under 30 minutes…..

Link to demo app here (this is not real data in case you’re wondering !): https://demo-medical-results-from-claude.netlify.app/

Did You Know? 

Till next time,

And The Rest Is Leadership