From Modems to Millions: The Accidental Birth of BoSacks’ Media Newsletter

If you’ve worked in publishing for any length of time, you’ve probably heard of Bo Sacks. And if you haven’t, well… where have you been hiding? His Heard on the Web: Media Intelligence newsletter has been landing in inboxes for decades, long before “email marketing” was even a term.

But what most people don’t know is that it all started with one AOL account, a squealing Hayes modem, and… no one to email.

AOL, Floppies, and the Birth of a List

It was the late ’80s, and Bo was production manager at Ziff-Davis, right in the middle of the magazine boom. Think bound-in inserts, blow-in cards, and enough promotional gimmicks to make a postal worker blush.

One day, a quirky project landed on his desk: attach floppy disks—yes, the breakable kind—to magazines. Pulling that off earned him a little thank-you gift from America Online: a free email account.

That might not sound like much now, but back then? It was like being handed the keys to a brand-new continent. The problem was, no one else lived there. Bo had exactly one person to write to—his old college roommate at Time Inc. So they traded notes about the arcana of magazine production: paper stocks, press checks, binding quirks.

A coworker joined the conversation. Then another. Before long, the list grew from two people to a few dozen. By 1992—before the modern web even existed—Bo’s impromptu email list had 1,000 subscribers. No ads. No growth hacking. No algorithms. Just word of mouth in publishing circles.

That accidental community is now the longest-running e-newsletter in the world, with over 16,750 subscribers in every corner of the media industry.

Why People Listened

Bo wasn’t just some guy with an email account. He’d been in publishing since 1970, starting with his own weekly newspaper in New York before moving into the alternative press and becoming an early leader at High Times magazine. Over the years, he held almost every job you can think of: publisher, editor, pressman, production director, senior sales manager—you name it, he’s done it.

His résumé reads like a map of the media world: McCall’s, Time Inc., New York Times Magazine Group, International Paper, Ziff-Davis, CMP, Bill Communications. He’s also co-founded mediaIDEAS, advised universities, and picked up more industry honors than some people collect coffee mugs—including a Publishing Hall of Fame induction and a Niche Media Lifetime Achievement Award.

In other words: when Bo talks about publishing, people tend to listen.

From Dial-Up to AI

Looking back, Bo sees a clear parallel between those early email days and today’s AI moment. Back then, the revolution was about access and speed—email leveled the playing field for independent voices. Today, it’s about intelligence and control—figuring out how to use powerful tools without losing the human voices that make content worth reading.

And his advice hasn’t changed much in 30 years: the winners aren’t the ones with the shiniest tools. They’re the ones who know why they’re using them—and have the stamina to keep producing meaningful work.

If you’ve never subscribed to Heard on the Web, it’s worth it just to see how a “Paleolithic” email experiment turned into a daily conversation that still connects publishing pros across the globe. And it’s also proof of something we tend to forget: sometimes the most enduring things start as an accident… followed by decades of showing up, day after day.

Hat tip to BoSacks’ newsletter for sharing this origin story. You can read the full piece here: BoSacks Speaks Out: From Modems to Millions – The Accidental Birth of a Newsletter

What Hearst’s AI Playbook Can Teach Smaller Newsrooms

I first came across this in the BoSacks newsletter. The original article — Hearst Newspapers leverages AI for a human-centred strategy by Paula Felps at INMA — lays out how Hearst is rolling out AI across its network.

Now, you might be thinking: “That’s great for a chain with San Francisco-based innovation teams and a dozen staffers dedicated to new tools… but what about us smaller or niche outlets that don’t have a DevHub?”

That’s exactly why this is worth paying attention to. Hearst’s approach isn’t just about expensive tech — it’s about structure, guardrails, and culture. Those translate no matter the newsroom size.

Hearst’s AI Guiding Principles

✅ What We Do

  • Embrace generative AI responsibly.
  • Stay aligned with Legal and leadership.
  • Involve newsrooms and journalists across the organization.
  • Create scalable tools that help journalists.
  • Keep humans deeply involved.

🚫 What We Don’t Do

  • Tarnish our brands for quick wins.
  • Mass-publish AI-generated slop.
  • Mislead our audience or avoid transparency.
  • Let bots run without oversight.
  • Do nothing out of fear of change.

Here’s the big picture:

  • Clear principles: They’ve drawn a hard line on what AI will and won’t do. It’s in writing. It’s shared. And everyone’s on the same page.
  • Human-first workflows: Every AI-assisted output gets human review. No shortcuts.
  • Small tools, big wins: Their AI isn’t all moonshots. Some of the biggest gains come from automating grunt work — things every newsroom wrestles with.

Why smaller newsrooms should take notes

  • You might not have a Slack-integrated bot like Hearst’s Producer-P, but you could set up a lightweight GPT workflow for headlines, SEO checks, or quick summaries.
  • You probably can’t scrape and transcribe every public meeting in the state, but you could start with one high-value local board or commission using free/cheap transcription paired with keyword alerts.
  • You might not launch a public-facing Chow Bot, but you could make a reader tool that solves one local pain point — from school board jargon busters to a property tax appeal explainer.

The secret here isn’t deep pockets — it’s intentional design. Hearst put thought into categories (digital production, news gathering, audience tools), built policies to match, and then trained their people. That part costs time, not millions.

As Tim O’Rourke of Hearst put it:

“We try to build around the expertise in our local newsrooms. That’s our value — not the tech.”

For smaller outlets, that’s the blueprint. Start with what you do best. Add AI where it can actually save time or uncover new reporting angles. Keep your humans in control. And make sure your audience always knows you value accuracy over speed.

Quick wins for small newsrooms

  • Write your own “What We Do / What We Don’t Do” AI policy in plain language.
  • Pick one workflow bottleneck and pilot an AI tool to tackle it.
  • Build an internal “AI tips” Slack channel or email chain to share wins and lessons.

You don’t need a DevHub to start. You just need a plan — and maybe the courage to experiment without losing sight of your values.

The Generative AI Paradox: How Small Businesses Can Win Without Tech Giant Budgets

There’s a fascinating New York Times article making the rounds about the “generative AI paradox” — the fact that corporate spending on AI is exploding, but the bottom-line payoff just isn’t there yet.

Big players like Microsoft, Amazon, and Google are raking in AI profits. Nvidia is selling chips like hotcakes. But for most companies, especially those outside the tech sector, AI is still in the “lots of money in, not much money out” stage. McKinsey says 80% of companies are using generative AI, but nearly the same number report no significant financial impact. That’s… sobering.

So if you’re a smaller firm, without billion-dollar budgets or a 60,000-person tech staff, how do you even begin to make AI work for you?

1. Think in Terms of Targeted Wins, Not Total Transformation

One big takeaway from the NYT piece is that the small wins are what’s sticking.

  • USAA uses AI to assist (not replace) call center staff.
  • Johnson Controls trims 10–15 minutes from repair jobs.
  • JPMorgan automates report drafting and data retrieval.

Notice what’s missing? Nobody’s saying, “AI replaced half our workforce and tripled our profits overnight.” These are micro-efficiencies that add up — and they’re exactly where smaller firms should focus.

For you, that might mean:

  • An AI tool that drafts first-pass proposals or reports.
  • Customer service chatbots that handle basic queries before a human steps in.
  • AI-powered search across your internal documents so your team stops reinventing the wheel.

2. Use the 80/20 Rule for Pilots

Here’s the trap smaller firms can fall into: thinking AI requires a big, complex rollout. It doesn’t. Start with one process where 80% of the work is repetitive and rules-based. Then, find an AI tool to chip away at that 80% — leaving your team to focus on the 20% that requires human judgment.

The NYT piece points out that 42% of AI pilot projects were abandoned last year. That’s not failure — it’s course correction. Shut down what’s not working quickly, but keep the lessons learned. Small firms can be nimbler than giants here, turning failed pilots into better second attempts.

3. Borrow the Big Guys’ Guardrails Without Their Bureaucracy

JPMorgan locked down security and data governance before expanding AI to 200,000 employees. Smaller firms can’t afford that scale, but you can still:

  • Keep sensitive data off public AI tools.
  • Choose platforms with strong privacy controls.
  • Train your team on what AI shouldn’t touch just as much as what it should.

The less you have to unwind later, the faster you can scale what works.

4. Invest in People Before Platforms

One overlooked point in the NYT reporting: many AI failures come from “human factors” — employee pushback, lack of skills, customer distrust. For a small business, a \$50,000 AI investment can be sunk in six months if the team doesn’t adopt it.

Sometimes the best first “AI spend” isn’t the tool — it’s the training. Even a few hours of workshops on prompt writing, tool selection, and workflow integration can double your ROI.

The Bottom Line
Small firms have a hidden advantage: agility. You’re not burdened by legacy systems, sprawling compliance departments, or shareholder expectations of quarterly AI miracles. You can try small, learn fast, and scale the wins that fit your business.

The generative AI paradox isn’t a reason to avoid AI — it’s a reason to approach it with focus and discipline. And in five years, when the tech giants are still “optimizing their AI stack,” you might already have a handful of well-oiled AI processes quietly making you more efficient every day.

Memorable Takeaway:
You don’t need a billion-dollar AI budget — you need a billion-dollar habit of finding and scaling the little wins.

Crime Is Dropping Fast — Here’s Why It Might Stay That Way

If you’ve been doomscrolling crime headlines lately, you might want to sit down for this: we’re actually living through one of the sharpest drops in violent crime in modern U.S. history. The murder rate? On track to hit a 65-year low in 2025. That’s not a typo.

I came across The New York Times Editorial Board’s piece, Crime Keeps Falling. Here’s Why, and it’s worth a pause before we all rush off to the next crisis. Their argument boils down to two big drivers — one social, one about policing — and together they paint a surprisingly hopeful picture.

The Social Glue Came Back

Remember 2020? Of course you do. Everything closed. Kids were stuck at home. People yelled at each other about masks, vaccines, and politics. The air felt tense, and not just because of the virus. Sociologists call it anomie — that breakdown of the subtle social rules that keep communities functioning.

When that frays, rules start to feel optional. We saw it in reckless driving, road rage, petty theft, even little things like talking through movies. Crime spiked.

Fast forward to today: schools are open, churches and community centers are running, people are back in public spaces. That “everyone-for-themselves” vibe has eased, and crime has fallen with it.

The Policing Pendulum Swung Back

During the 2020 protests, “defund the police” became a rallying cry. While most cities didn’t actually slash budgets, morale took a hit. Some officers quit, others pulled back, and enforcement of low-level crimes got… patchy. Parts of cities felt lawless.

Now? Staffing levels are stabilizing, enforcement has ticked back up, and some states have rolled back overly lenient policies. The defund movement is widely seen as a failed experiment, and there’s been a shift back toward visible policing — without throwing out reform efforts entirely.

Why This Matters

The Times warns against complacency. Even with today’s numbers, violent crime in the U.S. is still higher than in most peer countries. Lax gun laws remain a glaring weak spot. And while reforming abusive policing practices is essential, gutting basic law enforcement isn’t the way to do it.

Their takeaway is refreshingly simple: social trust matters, and so does effective law enforcement. When both are in good shape, the crime needle moves faster than we think.

My Takeaways for Leaders and Policymakers

  • Keep the social infrastructure strong — schools, parks, and civic institutions are as much crime prevention tools as any patrol car.
  • Balance reform and enforcement — fix what’s broken in policing without abandoning the basics.
  • Watch the trends early — use data to see the next spike before it hits.

💡 Memorable line from the editorial: “Law enforcement matters, and the national mood matters — and together they can move the crime needle faster than we often believe.”

If you’re in public policy, policing, or community leadership, bookmark this one. And if you’re just a citizen trying to make sense of the headlines, here’s some rare good news: the numbers are moving in the right direction — and we have a decent idea of why.

Daily Links: Thursday, Aug 14th, 2025

In my latest blog post, I dive into OMNARA, a game-changing tool for managing AI agents. Created by experienced engineers from Meta, Microsoft, and Amazon, and backed by Y Combinator, OMNARA offers a unified dashboard to monitor, manage, and collaborate with your AI fleet. Check it out to learn how it can streamline your AI operations!

  • OMNARA: Monitor, manage, and collaborate with your entire fleet of AI agents from a single, unified dashboard. Built by ex-AI engineers from Meta, Microsoft, and Amazon. Y Combinator backed.

AI in the Newsroom: Why It Should Be Your Smartest Intern, Not Your Star Reporter

Practical AI tools and governance tips for small and niche newsrooms that want smarter reporting, not robot reporters.

If you’ve been anywhere near a journalism conference in the past year, you’ve probably heard the AI hype: “It’s going to replace reporters.” “It’s the future of investigative journalism.” “It’s going to write all our stories for us.”

But here’s the reality check, courtesy of journalist-technologist Jaemark Tordecilla — someone who’s actually been in the trenches building AI for newsrooms. In a recent INMA piece, Tordecilla put it plainly: AI is a terrible journalist. It doesn’t chase leads, smell a rat, or spot the story between the lines. What it does do exceptionally well is the grunt work — the sifting, sorting, and summarizing that lets you get to the important stuff faster.

And that’s the mental shift small and niche news organizations need to make: stop asking AI to be the reporter, and start asking it to make your reporters’ jobs easier.

Tools That Complement, Not Replace, Human Skill

If you’re running a small newsroom with limited staff, think of AI as your hyper-efficient intern — one that doesn’t sleep, doesn’t take lunch breaks, and doesn’t mind doing the boring bits.

Here are a few practical tools you could build or adopt:

  • Data Sifters
    AI models that can ingest giant PDF reports, meeting transcripts, or spreadsheets and spit out bullet-point summaries or proposed headlines. Your reporter glances at the output and decides if it’s worth a deeper dive.
  • Budget Chatbots
    Exactly like Tordecilla’s tool for “chatting” with the Philippines’ 700,000-line national budget. For local publishers, this could mean feeding your city or county budget into an AI tool and asking questions like: How much did we spend on police overtime last year? or Which departments’ budgets increased the most?
  • Pattern Spotters
    Tools that flag anomalies or trends in datasets — e.g., tracking how often a government department awards contracts to the same vendor, or how property sales spike in certain neighborhoods.
  • Fast-Format Converters
    AI-assisted workflows that can take a long-form investigative article and quickly produce a podcast script, social video captions, or illustrated explainers. The key: these formats should be reviewed and fine-tuned by humans before publishing.

The Governance Question: Who’s Driving This Thing?

If AI is going to become part of your newsroom’s workflow, you need rules of the road. For small and niche publishers, governance doesn’t have to be a 40-page corporate policy, but it does need to answer some core questions:

  • Transparency: Will you disclose when AI is used in research, production, or content creation? How?
  • Attribution: Who “owns” AI-generated outputs in your newsroom — and how do you credit sources if AI pulls from third-party data?
  • Bias Checks: How will you review AI-generated summaries or insights for skew, especially when dealing with politically sensitive topics?
  • Ethical Boundaries: Where will you not use AI? (For example, generating deepfake-like images of people, or creating composite quotes.)
  • Review Protocol: Who signs off on AI-assisted work before it goes public? Even small teams should have a second set of eyes on anything AI touches.

A lightweight governance structure might be as simple as a one-page “AI Use Policy” taped to the newsroom wall. The important part is that everyone knows the rules — and follows them.

Why This Matters for Small Newsrooms

Big national outlets can afford to burn cycles experimenting with AI. You probably can’t. That’s why your AI playbook should focus on high-leverage tasks: the work that’s essential but time-consuming, where AI can give you a multiplier effect without compromising your credibility.

The payoff? More time for your reporters to be out in the community, making calls, filing FOIA requests, and doing the human work AI can’t touch.

Memorable Takeaway:
“AI is good at finding patterns in data; humans are good at finding meaning in those patterns. Keep it that way.”

Daily Links: Wednesday, Aug 13th, 2025

In this post, I explore some intriguing links I’ve been reading. You’ll find insights on Apple’s upcoming Siri updates and supply-chain strategies courtesy of Bloomberg. There’s also a fantastic free guide on improving your social skills. If you’re a tech enthusiast, you’ll love how Claude Code enhances my work and fun. Plus, discover how my quest for the perfect to-do app led me back to the simplicity of a .txt file!

Stop Arguing, Start Asking: Why Prompt Literacy is the Next Universal Skill

AI isn’t magic. It’s not malicious. It’s not even confused.
It’s a tool — and like any tool, what you get out of it depends on how you use it.

Two recent takes — Linda Ruth’s Stop Arguing with AI: Prompting for Power in the Publishing World and Kelvin Chan’s AP piece One Tech Tip: Get the most out of ChatGPT and other AI chatbots with better prompts — arrive at the same destination from different roads. One speaks to editors and publishers; the other to everyday AI users. But the core message is identical: the quality of your AI output starts and ends with the quality of your input. (Hat tip to the always-essential BoSacks newsletter where I first spotted both articles.)

From the Newsroom to Your Laptop: Same Rule, Different Context

Ruth frames AI as part of a publishing professional’s toolkit — right up there with headline writing and layout design. If you ask a model for feedback on a manuscript without providing the manuscript, expect nonsense in return. It’s like asking a book reviewer to critique a novel they haven’t read.

Chan’s advice mirrors this in broader strokes: skip vague prompts, give clear goals, and feed the model context and constraints. Add personas to shape tone, specify your audience, and don’t be afraid to iterate. The first prompt is rarely the last.

The Practitioner’s Mindset

Whether you’re an editor, marketer, small business owner, or teacher, three habits will instantly improve your AI game:

  1. Provide context — the more background you give, the better the results.
  2. Set constraints — word count, format, style — so you get something usable.
  3. Iterate — treat AI as a collaborator, not a vending machine.

Think of AI as a “brilliant but distractible employee”: give it structure, keep it focused, and check its work.

The Bigger Picture

The skeptic will say this is common sense — ask better questions, get better answers — and they’re right. But prompt literacy is becoming a baseline skill, much like search literacy was twenty years ago. The contrarian might argue AI should adapt to us, not the other way around. The systems thinker sees a familiar pattern: early adopters learn the machine’s language, then the tools evolve until the complexity disappears behind the scenes.

Until that happens, prompt engineering is the bridge between what AI can do and what it will actually do for you.

Turn questions into results. Don’t just wonder what AI can do — start guiding it. Download my free, printable AI Prompt Quick Guide for proven prompt formulas you can use today.

Action Steps You Can Use Today

  • Create a personal or team prompt library for recurring tasks.
  • Refine in conversation — don’t settle for the first draft.
  • Experiment with personas and audiences to see how the output shifts.
  • Always verify — a polished answer can still be wrong.

In short: Master the prompt, master the tool — and in mastering the tool, you expand your reach.

Crisis Mode: Why Small Publishers Should See Opportunity in the Chaos

Every now and then, I read something from a media veteran that feels like it’s aimed right at the big players — but still lands squarely in the lap of small and niche publishers. That’s exactly what happened with Chris Duncan’s upcoming keynote at the FIPP World Media Congress.

Duncan’s career is full of steering through storms — launching The Times on the iPad (when that was brand new territory), leading through COVID, and now advising on the AI tidal wave that’s hitting every corner of publishing. His core message? Publishing thrives in crisis.

Now, “thrives” might feel like a stretch if you’re running a three-person operation and trying to keep the lights on. But here’s where the small guys might actually have an edge: when the ground shifts under everyone, agility beats scale.

What small publishers should take away

1. AI isn’t just a newsroom curiosity — it’s a traffic problem.
Yes, AI tools can help you cut costs and automate grunt work. But Duncan’s warning is clear: generative AI could cut off more referral traffic than Google already has. For small publishers, that means you can’t afford to be a “search-dependent” business. Your audience has to remember you and seek you out.

2. Innovation isn’t optional.
He’s blunt: mobile journalism hasn’t seen much truly new since about 2012. That’s both sobering and exciting. If you’re a niche publisher, you don’t need to outspend The New York Times — you need to outthink them in your lane. That might mean interactive features, audio companions to your stories, or even an “insider’s app” for your core audience.

3. The platform era is shifting — be ready.
Duncan thinks we’re past the peak of Google and Meta’s dominance. That’s a rare window to build distribution without depending entirely on them. When big platforms are distracted by regulators and market shifts, you can make a move to deepen your direct audience connections.

Where to put your focus next

Here are three action items I think every small or niche publisher should put on their whiteboard after reading Duncan’s comments:

  1. Build direct audience pipelines.
    Start or double down on newsletters, podcasts, private communities, or events. Make sure your readers’ path to your content doesn’t depend on an algorithm.
  2. Test one “genuinely new” product feature in the next year.
    Could be a micro-app, an interactive archive, or a new storytelling format. The goal is to prove you can innovate without waiting for the industry to hand you a playbook.
  3. Scenario-plan for a search traffic cliff.
    If your Google referrals dropped 50% tomorrow, how would you adapt? Do that planning now while you have the luxury of time.

Duncan’s not saying this will be easy — far from it. But he is saying that urgency forces experimentation, and experimentation is where breakthroughs happen. For small publishers, the trick is to use your speed, focus, and audience intimacy as weapons in this fight.

You may not have a “war room” of strategists, but you do have something the giants often lack: a direct line to a loyal audience that cares deeply about your coverage. That’s your moat. Guard it, grow it, and use this crisis moment to get a little scrappy.

If you want the full keynote preview, it’s worth a read: Publishers work best in some form of crisis.

Takeaway for the fridge:
Crisis is coming. The question is — will you let it happen to you, or will you make it work for you?

Daily Links: Tuesday, Aug 12th, 2025

In my latest blog post, I dive into the architectural advancements from GPT-2 to gpt-oss and compare them with Qwen3. I also spotlight “bolt,” a high-performance, real-time optimized, statically typed embedded language written in C. It’s an exciting exploration into cutting-edge AI and coding technologies that I think you’ll enjoy!