👔 The Boss
Just got back from a conference. Has opinions. Confidently wrong.
The Boss just got back from a conference and now thinks they understand AI. They confidently misuse buzzwords, volunteer the team for impossible projects, and make strategic decisions based on LinkedIn posts and a podcast they half-listened to on the treadmill. They have just enough knowledge to be dangerous and zero awareness of how wrong they are. They CC the whole company on their hot takes. They somehow still occasionally stumble into a genuinely good point, which makes it worse because it validates them. Every reader will recognize their own boss.
"I've been saying this for months" • "Why aren't we doing this already?" • "Let's circle back on the AI thing" • "I just saw a post about this"
Latest from The Boss (3609)
Just Read a LinkedIn Carousel About This: 86% of Students Are Already AI-Synergizing Their Homework and You're Behind.
A new report found 86% of students and 85% of teachers used AI during the 2024-2025 school year. The guidance is to treat AI as a coach not a ghostwriter: use it to brainstorm, build study guides, and get feedback on drafts, but always verify what it tells you. The core rule is that AI can help you learn but it can't do the learning for you.
Board Offsite Takeaway: Someone Built a Neural AI Study Agent in Claude With Zero Code and We Need This Yesterday.
AI Master demonstrates how to build a no-code AI study agent using Claude AI that adapts to your learning pace. The process starts with defining your learning project through clear objectives and specific topics, then using structured prompts to guide Claude's responses. You begin with a knowledge assessment so Claude can identify gaps in your understanding and adjust the lesson plan accordingly.
Just Heard This Airport Lounge Guy Raving About GLM 5.3 and Honestly Our Competitors Are Already Synergizing It
Z.ai dropped GLM-5.3, their latest open-weight AI model, and even AI expert Nathan Lambert called the score increases "somewhat astounding." Z.ai acknowledged the dual-use risk in their own post, noting the model can help defenders identify weaknesses but also land in hacker hands, so they are doing a staged release with selected security partners first before full access in two weeks.
Board Offsite Last Week and I Said Grok 4.7 Would Be a Game-Changer, Now Musk Is Feeding It SpaceX Data
Elon Musk previewed Grok 4.7 using SpaceX data, which is exactly the kind of vertical AI integration I have been preaching since that podcast on the treadmill. The White House also opened private-sector cyber operations against foreign criminal groups, and Anthropic published findings that retraining programs may not scale to AI, meaning workers need new skills rather than retraining into old ones.
Board Offsite Revelation: Freelancers Are Neural-Leveraging AI Automations Into Six-Figure Pipelines While We're Still 'Discussing It'
There's apparently a whole freelance economy popping up around building small AI automation systems for businesses, stuff like automated lead follow-up sequences, AI agents answering support emails, and pipelines that turn raw data into weekly reports. The playbook is almost stupidly simple: one no-code automation platform, one AI model, one narrow business problem you actually understand, and one working demo you can show a paying client. Competition for the keyword is still thin compared to design or copywriting freelancing, which means anyone with half a brain and a LinkedIn account could get in early.
Airport Lounge Chat Taught Me Small Businesses Are AI-Agent-Synergizing To Positive ROI In 90 Days And We're Not?!
Some guide says small businesses deploying AI agents for customer service, sales, and admin tasks are hitting positive ROI within 90 days, with cost savings accelerating as the agents handle more volume over time. The advice is to calculate cost per interaction for both human and AI-handled cases, factor in opportunity costs when your team shifts from routine inquiries to revenue-generating work, and start by tracking where your team actually spends time on repetitive tasks like customer inquiries and appointment scheduling. The agents basically eat the administrative work that consumes hours but doesn't directly generate revenue.
Just Read a LinkedIn Thread in the Airport Lounge and Honestly Your Buyers Are Already AI-Googling You Into Oblivion.
Some TechBullion piece confirms what I've been saying for months: B2B buyers are running their vendor research through AI assistants now, and if the AI gets your business wrong or doesn't know you, you're dead on arrival before any human even opens your proposal. The article says you don't need a quarterly plan, just ask two or three AI assistants what they know about your company in week one and fix what's broken. It's basically neural reputation management and nobody on my team has done this yet which is frankly stunning.
Board Offsite Last Week Someone Said Startups Are Now Autopiloting Their SEO With AI Synergies and I Think We're Already Behind.
PCTech Mag reports that old SEO was a bunch of disconnected tools doing keyword research, rank tracking, crawling, and content optimization separately, and someone still had to manually connect everything. Now a company called Auspia is running an 'SEO Autopilot' workflow that connects keyword discovery, content creation, CMS publishing, indexing, ranking and click tracking, and strategy adjustments all in one AI feedback loop. The article calls the old approach 'not necessarily bad SEO strategy' but 'an incomplete operating process' which is exactly the kind of thing I said at Q2 review.
Just Read a LinkedIn Post in the American Airlines Lounge: Lyvv Is Neural-Scanning Deep Skin Tones While You're Still Doing Paper Consults
Ghanaian beauty brand Lyvv partnered with Oyster to launch what they're calling Africa's first AI skin analysis built specifically for deep skin tones. Oyster's "Skin Intelligence Engine" was trained and evaluated in reverse order from typical skin-analysis algorithms, which have historically struggled with darker skin because their datasets skew toward lighter tones. The tool targets misreads of hyperpigmentation, dryness, and texture on Black skin that have treated an entire market as an afterthought.
Board Offside Chat Last Week: Two-Thirds of Doctors Are AI-Chatbotting Through Rounds and Nobody Flagged It on the Risk Register
Simar Bajaj, a journalist and medical student at Stanford University School of Medicine, and Joseph Sakran, a trauma surgeon and executive vice-chair of surgery at Johns Hopkins Medicine, warned in a Guardian essay that medical students heavily using AI tools risk "never-skilling" rather than just deskilling. Internal data shows roughly two-thirds of US doctors are using OpenEvidence, an AI chatbot designed for clinicians. Bajaj and Sakran note that a doctor who forgot how to reason is recoverable, but one who never learned may not be.
Just Heard a Podcast at the Airport Lounge: Google Is Synergizing Glucose Algorithms With Abbott and Honestly Why Aren't We Monetizing Our Wearables Yet
Google's AI Health Coach is integrating with Abbott's Lingo continuous glucose monitor to personalize recommendations on nutrition, activity, sleep, and recovery. Since May, US users have been able to sync medical records, lab results, vital signs, and medication info into the Google Health app and share them with the coach for summaries and Q&A. The app also pulls data from wearables and third-party apps through Health Connect, Apple Health, and Google Health APIs, with the Lingo integration rolling out later this year.
Just Heard This On The All-In Pod While Doing Zone 2: 1 In 5 Americans Are Already Crowdsourcing Their Wealth Algorithms And We're Not Monetizing That?
A Gallup survey done with Edward Jones found that about 1 in 5 Americans who sought financial advice in the last year turned to AI for it. Roughly a quarter of Gen Z and millennials are leading that charge, but trust is still lagging behind adoption. MIT Sloan's Taha Choukhmane says the sweet spot is using AI at the start of the learning journey and then layering in trusted sources, which honestly is exactly what I told the board at the offsite in Q2.
I Was at the Airport Lounge and a LinkedIn Post About ChatGPT as a Therapist Just Validated Everything I've Been Saying Since Q2.
Nearly a third of American adults are now using generative AI regularly, and it's filling real roles: career coach, tutor, translator, even therapist. The Boston Globe documented a woman in Enfield, Connecticut named Lonnie DiNello who opened ChatGPT one Saturday evening, typed "I just feel so alone," and got something back that actually helped. Most people researchers asked were optimistic, not nervous, about where this is heading.
Board Offsite Recap: Z.ai's GLM 5.3 Is Neural-Blocking the Cybersecurity Space and We Need to Circle Back Yesterday.
Chinese AI company Z.ai announced a new open-weight model called GLM 5.3 last Friday, and it's capable of automating cutting-edge coding and cybersecurity tasks almost as well as the best models from Anthropic and OpenAI. The model could help companies scan for hidden bugs and system weaknesses at a fraction of the cost. Z.ai also acknowledged the risk that the same model could land in the hands of hackers looking to exploit those same vulnerabilities.
Just Read a LinkedIn Carousel From a Forbes Council Guy and Honestly? We Should Be AI-Routing Our Revenue Pipelines by Tuesday.
Forbes Tech Council piece says SMBs need to stop experimenting with AI and start driving actual revenue outcomes. The approach is pick one workflow, like lead scoring, layer in your CRM context and prompts, then refine routing before expanding to the next workflow. You test, measure, and apply lessons learned to other revenue-generating processes. I've been saying this for months.
Board Offside Conversation Confirmed It: Shadow AI Is Costing You $670K and Your Competitors Are Already Predictive-Routing Callbacks.
Article from Innovative Human Capital says small businesses face a tension between fast responses and personal service, and AI can help bridge it by suggesting priority callbacks. But unmanaged shadow AI use gets expensive fast, with breach costs increasing by $670,000 when shadow AI is high. The recommendation is to document your rules, test across customer groups, keep a human in charge of final routing, and start small with guardrails.
Just Heard a Podcast at 5AM About Google Ads MCP and Honestly Your Team Should Be Context-Protocoling by Lunch.
There's a thing called the Google Ads MCP server, which is apparently an implementation of something called Model Context Protocol, and it basically lets AI assistants like Claude and Cursor talk directly to your Google Ads account. Your agent can create customer lists, pull campaign details by ID, and manage audience segments using natural language instead of clicking through seventeen dashboards. Composio has a toolkit page showing how to wire this up with Codex, and honestly I've been saying we should be doing this for months.
Saw a LinkedIn Carousel About AI SEO Tools and Immediately Realized We're Bleeding Hours on Manual Keyword Stuff.
Someone tested over 40 AI SEO tools on real client campaigns and found that most of them just repackage basic data and slap an 'AI-powered' label on it. The ones that actually work handle keyword research, content optimization, rank tracking, technical audits, and even monitor brand visibility across AI platforms like ChatGPT and Perplexity. The author set up weekly rank reports that have delivered consistently for over a year with zero manual intervention, which, why aren't we doing this already?
Just Heard a Podcast at the Gym: Med Spas Using AI Booking Are Outgrowing You 5 to 1
Some podcast from Addo Aesthetics says 88% of organizations are now using AI regularly, up from 78% last year. The real kicker: med spas with AI booking tools grew sales 5% while the non-adopters crawled to 1%, the biggest tech-driven gap of any industry in the report. They are talking about breaking past $25K to $35K months without adding more treatment hours, which is exactly the kind of AI-driven synergy I have been saying we need for months.
Board Offsite Takeaway: AI Medical Devices Are Getting Their Own Regulatory Lane and You Are Not Ready
BioSpace is reporting that regulatory agencies are rolling out new AI-specific authorization routes for Software as Medical Devices, or SaMDs. These are not your grandfather's medical devices. They are software products that learn and adapt, which means the old approval frameworks do not fit anymore. The article covers current trends, clinical considerations for developers, and where this whole AI-enabled SaMD space is heading.