The MRKT3.0 Editorial team submit an entirely fictional account of Jon, the Prompt Bro.
November 2025
Jon had been watching Lex Fridman on YouTube again. He had spent hours following what users had been posting and commenting in AI & Tech groups on LinkedIn and Reddit. He even checked X occasionally as he liked and reshared Dario Amodei’s posts at least once a week for months, but he hadn’t actually used Claude properly yet.
At least a few of his friends had ‘accepted’ (at least he assumed that they’d accepted) that he had become a ‘Prompt Bro’. If he had been over-sensitive he may have suspected that they were laughing at him. But they had clearly accepted his new-found enthusiasm for everything AI. They obviously respected how smart he must be to know all the things he did about large language models and agentic workflows in general.
Jon had worked hard to build a reputation as a thoughtful professional. His Dad had always told him to play it safe. And with AI moving so fast, he still worried about whether he might look foolish if he finally committed to one tool. He had colleagues and a LinkedIn following to think about. He’d always been known as a careful early adopter. Everybody knew that Jon was good with technology.
Jon had heard from people like Dario Amodei, that serious-looking guy from that Anthropic company everyone kept writing about, that Claude was fundamentally different from other AI models. Something about being genuinely helpful and safe at the same time. Surely it was still early enough to get ahead of the curve?
Jon had mapped the AI landscape across three whiteboards. He was really proud of all the capability comparisons he’d managed to chart. He had finally created a Claude.ai account and was already composing his first LinkedIn post about it. His colleagues would think he was so ahead of the curve.
Claude ends November just out of reach. Jon has opened the interface exactly once, typed “hello”, been genuinely impressed by the response, and then closed the tab to do more research first.
December 2025
Jon had been deep in AI & Tech groups on LinkedIn and Reddit again. Jon had been impressed with how seriously Anthropic was taking AI safety, but more importantly someone at his company (probably Emily) had quietly built an internal Claude integration, and the whole team was already using it over the Christmas party planning calls!
Jon had been very steadfast in resisting the temptation of fully committing to Claude, it wasn’t as cool as ChatGPT surely. But he’d finally gone all-in after seeing what had happened to his colleague’s output after she’d started using it daily. All his friends thought being efficient was cool. He’d already been showing off his first AI-generated strategy deck — it was his favourite thing to do in December. It made him look sharp, and helped gloss over the fact that he’d only started once everyone else in the office already had.
His manager and colleagues would be so proud of him when his output started to quadruple in 2026. He had already started to think about what to automate first. The weekly status reports and the end-of-year competitor review to start with.
Had he gone deep enough, though? Perhaps he could build a custom Claude Project, or even try the API? Otherwise, he’d lose out on the real gains, wouldn’t he? His colleagues would think he was so forward-thinking when he showed them what he’d built. He was so innovative.
Somehow, he’d missed this post on X from a few days before:
Claude ends December fully installed on Jon’s laptop, his phone, and pinned in his browser toolbar. He has used it to rewrite four emails and is already telling people he “works with Claude AI daily.”
January 2026
Jon had been mainlining Lex Fridman episodes and AI newsletters. He’d signed up to something called an “AI newsletter stack” because someone on a podcast had mentioned it was what serious people read. Jon had been learning about Anthropic’s rivals from the community: the OpenAI crowd, the Google DeepMind crowd, even this European Mistral crowd. He’d never really compared them head-to-head before. Must be the thing to do now that everyone had opinions on benchmarks and the EU AI Act enforcement timeline.
Jon had been at a New Year networking event showing people his Claude Projects setup on his phone. It was so much more impressive than the basic ChatGPT setup his colleague used or the ChatGPT integration that most of his friends had settled for. Everything from San Francisco was always so much more considered. Jon was almost overcome with his own wonderfulness.
Jon had set up a Claude Project for every area of his professional life in preparation for what he was calling his “AI-first 2026.” He would be one step ahead and show all his colleagues how much more productive than them he was. His colleagues had also been asking him more questions about AI. It was obviously just because he was so much better at this than they were.
January was the first time that Jon had managed to produce something genuinely impressive using Claude. His director was even showing signs of interest in what he was doing. But Jon hadn’t told her exactly how long the prompting had actually taken him yet. He would surprise her with a proper case study one day instead.
Claude ends January with Jon officially becoming a #Prompter. He has saved 51 minutes this month and told at least fourteen people about it.
February 2026
Jon had been doom-scrolling AI agent hype on X and LinkedIn. He’d been learning all about AI agents after a breathless article he’d read in Wired about “the end of manual work.” He’d even updated his LinkedIn headline to include the phrase “AI-augmented strategist”. Something that only proper thought leaders were doing.
Jon had also upgraded to Claude Pro. He’d heard about the extended context window and the new agentic capabilities rolling out across enterprise teams. It all sounded very good for his personal brand, and so many of his colleagues were looking at him enviously. Obviously.
He’d been showing off his new knowledge of prompt engineering after reading a post about “chain-of-thought reasoning” and “extended thinking mode.” When he explained to his much less informed colleagues what it meant, it was so obvious that they all admired him for how technically sophisticated he’d become.
His wife had been really supportive recently too. She’d even offered to listen to his podcast episode idea: “AI Unlocked with Jon”, where he would talk about AI all the time, obviously. She was clearly proud of Jon’s vision and was showing him how much she appreciated his entrepreneurial spirit. Wait till she found out how he’d automated their holiday planning and how they’d save three whole hours booking a much better summer trip than that city break she’d been suggesting.
Jon had been learning more about all these #Prompters in February too. The end of the month was a bit nail-biting when Claude gave him a confidently wrong answer about a competitor and he’d already sent the analysis to his director. But none of his friends had noticed. Probably. Best not say too much about hallucinations anymore. Have you seen Severance Season 3 on Apple TV? The production design is genuinely extraordinary in spring.
Claude ends February with Jon having quietly deleted two LinkedIn posts after the comments section became complicated. He remains bullish.
March 2026
Jon had been telling himself he was “staying informed” on LinkedIn and Reddit again. He had spent a few hours following what users were posting in the usual AI & Tech groups as the whole Anthropic-Pentagon drama exploded. Apparently that man in the White House who all of Jon’s friends loved to hate had ordered federal agencies to stop using Claude after Anthropic refused to remove the model’s safety guardrails for fully autonomous weapons and mass surveillance. The Pentagon had even labelled them a “supply chain risk.”
Jon had read the entire MRKT3.0 piece on it and was now confidently explaining the “nuanced AI ethics and national-security implications” in comments in his groups. He’d been experimenting with Claude Code after Anthropic’s big update too, convinced it would finally revolutionise his technical workflows and make him look like the most agentic guy in the office.
He hadn’t told anyone about the experiment in the end. Things had turned awkward at work… His director was now asking for hard numbers on his “AI-first transformation.” All that supportive nodding in February suddenly felt tactical. Jon was scrambling to find a different success metric for the Q1 all-hands.
The European AI Act hadn’t helped either. Opinions in the groups were split down the middle, yet Jon had built a minor reputation for confidently explaining compliance implications. At least the newer members seemed impressed. He had joined when the biggest group had fewer than a thousand members. Now it was over 13,000. Jon was basically an influencer.
Claude ends March with Jon realising that AI can be confidently wrong as well as impressively right. And that even the companies building it could get dragged into political wars. He decides to keep #Prompting anyway.
April 2026
Jon had been quietly avoiding Lex Fridman for a few weeks. The AI groups were still active, but he was spending less time there. Someone had posted about how advanced models could theoretically hack almost any system, and the thread had spiralled.
His director had finally asked for a proper ROI report on all the Claude usage. Jon was now drafting a sleek four-slide deck, running every paragraph through Claude for polish, and hoping the next available slot in her calendar would buy him enough time.
He still believed the colleagues would see him as the most forward-thinking person in the room when he presented his “AI Adoption Framework” at the next town hall. They’d be queuing up for his prompt library. Surely.
Claude looked like it was genuinely transforming his output after a rocky start. And with rumours swirling about Anthropic’s new Mythos model breaking benchmarks left and right, Jon was already preparing his victory narrative.
Claude ends April with Jon still bullish, quietly deleting another LinkedIn post that hadn’t aged well, fully convinced the next model update would make him look like a genius.
#Prompter #ClaudeAI #AITools #ArtificialIntelligence #MRKT30 #AIWorkplace #MRKT3Summit #Anthropic #AIProductivity #EuropeanAI
Disclaimer
The views and opinions expressed in the story about Jon are completely random and not based on any particular individual or organisation.
The Editorial Team at #MRKT3.0 have taken all precautions to ensure that no persons or organisations have been adversely affected or offered any sort of career advice in this article.
This article is most definitely not Professional Advice.
Our readers are reminded that adopting AI tools can mean that you will embarrass yourself in front of your director, quietly delete LinkedIn and Reddit posts, and still not automate the thing you actually needed to automate in the first place.
