2025-08-11 [Insert Fantastical GPT-5 Headline Here]
ChatGPT-5 released this weekend (though I’m sure you already know this) and it’s highlighted an interesting divide in the world of AI users. On the tech side, there are headlines hailing it as a major move forward, one step closer to AGI, PhD-level expert, yadda yadda. On the power and casual user side, folks are mad that GPT-5 lost some of its personality. A snippet from the ChatGPT Subreddit titled “GPT-5 is a disaster” highlights how both sides are right:
GPT-4o had this… warmth. It was witty, creative, and surprisingly personal, like talking to someone who got you. It didn’t just spit out answers; it felt like it listened.
Now? Everything’s so… sterile. Formal. Like I’m interacting with a corporate manual instead of the quirky, imaginative AI I used to love. Stories used to flow with personality, advice felt thoughtful, and even casual chats had charm. Now it’s all polished, clipped, and weirdly impersonal, like every other AI out there.
Hmmm, like talking to PhD-level expert 🤔(no shade intended toward those with PhDs).
The varied responses shine light on what each group wants AI to be. Standard users want their AI assistant to be consistent and reliable with some warmth and personality (like H.E.R.). Tech-focused users want it to be smarter, faster, and more efficient (nearing HAL territory).
Wherever you sit on the line (maybe you don’t even use ChatGPT), it’s interesting to see how a tool so ingrained in the day-to-day can trigger these reactions.
We’re in an age where vibe-coded vaporware is going to get looks from top VCs because of the possibilities. All of the products on ProductHunt have some sort of AI capability. Companies are scrambling to stay relevant the only way they know how, jumping on the hype train and embedding AI into all of their products regardless of whether or not its a good idea.
Call me old-fashioned, but I don’t think I need an AI-powered refrigerator to tell me what’s inside my refrigerator. Instead, I’ll just…open…the door?
On the flip side, as the Head of Engineering for a Conversational AI team, I focus on improving operational efficiency and delivering more value to the company through AI use. It’s a fascinating world out there, one that’s rapidly evolving and it’s tough to know what tools and trends will have staying power.
My advice on how to proceed with the hype is to try it for yourself. Figure out what the repetitive tasks are that you wish you didn’t have to do anymore, and think of a way to offload to AI. What do you wish you could spend more time on? How might AI assist you in getting more time? You could even start with a basic prompt like:
You are an automation expert for casual users. I want to spend more time painting miniatures, but I have X, Y, and Z tasks that take up too much of my day. How can I automate these tasks without a budget? If you’re looking for some reading on AI and still forming your opinions on whether it’s good, bad, or both, I’ll aim to have an AI section in this newsletter every week, along with my standard management and engineering leadership links.
🧭 Engineering Leadership & Management
How Netflix, Stripe, and GitLab Built High-Performing Engineering Teams [Substack]
Three elite orgs. Three playbooks for building engineering excellence. Talent density, decision-making frameworks, and cultural clarity. Steal shamelessly.20 Powerful Leadership Quotes That Will Transform Your Thinking [Medium]
Not corporate fluff. These are quotes worth putting in board decks and taping to your fridge.The State of Engineering Leadership [Eng Leadership Newsletter]
A high-level scan of today’s leadership climate. Retention, alignment, and where execs are quietly panicking.Top Performer Who Has Lost Faith in You [Reddit]
When a high performer disengages, your leverage drops. Practical ways to rebuild trust without begging.When Team Structure Collides with Role Alignment [Medium]
Organizational Jenga. Reporting lines and actual responsibilities out of sync. Fix it before velocity crumbles.
🌍 Remote Work & Organizational Systems
Choosing Where to Spend My Team’s Effort [frederickvanbrabant.com]
A clear framework for deciding what work truly deserves your team’s energy.
🤖 AI in Engineering Workflows
Mastering AI at FAANG: A Roadmap from Junior to Senior Engineer [Medium]
A career progression cheat sheet for AI-native engineers. Skills, tools, and mindset shifts that matter.AI Acceptance Rate Is an Easy Measure to Spot Misuse [LinkedIn]
Laura Tacho introduces a KPI to catch AI misuse early. Keep code quality intact.Andrew Ng Calls Vibe Coding an Engineering Apocalypse [Perplexity]
Ng warns that vague prompts plus AI lead to sloppy, hard-to-maintain code.Vibe Coding and the Coming Engineering Apocalypse [WIRED]
Journalistic view of how AI-generated code could erode engineering discipline.Coding Agents Cross a Chasm [Singleton Blog]
AI coding agents are entering production use. Progress is real but integration is messy.Seizing the Agentic AI Advantage [McKinsey]
A structured guide to deploying AI agents as strategic assets.
📊 Prioritization, Strategy & Metrics
2025 Compensation Survey Report [SSUSA]
Fresh salary benchmarks for engineering, product, and data roles.Default Alive or Default Dead [PaulGraham.com]
The simplest way to measure startup survivability. Ignore it at your peril.
🔮 Broader Industry & Thought Leadership
Disney and Universal Sue Midjourney Over Copyright Infringement [The Verge]
A landmark IP case that will influence how generative AI trains on content.Can Your Brain Talk to Others While Asleep [Science Focus]
Early experiments suggest dream-state communication might be possible.How to Avoid the Ethical Nightmares of Emerging Technology [HBR]
A practical framework for responsible innovation.The Erosion of Quality in Large Language Models [ScienceDirect]
Research shows LLM quality declines without human oversight.