AI Side Hustles Half-true — works only if you do the unspoken work

By the Income Reality Check Editorial Team · Updated 2026-05-28 · 7 min read

Dan Martell’s 16 Claude hacks: what works, what’s beta, what costs extra

Verdict: Half-true — works only if you do the unspoken work. The features exist, but the smooth tour skips paid-plan gates, beta status, and security caveats Anthropic itself publishes.

Dan Martell’s 16-minute YouTube video, “Learn 97% of Claude in Under 16 Minutes,” has pulled in more than 612,000 views by promising a tour of every Claude feature worth knowing. He runs through 16 hacks — Gmail and Calendar connectors, Artifacts, Projects, Voice mode, the Chrome extension, Cowork, Claude Code, Skills, Claude Design — and frames each as a free productivity unlock. Most of the features are real. The trouble is what he glides past.

What the video actually claims

Martell runs Martell Ventures and previously founded SaaS Academy, the coaching outfit that was acquired by Precision in February 2026. His pitch in this video is not a dollar promise. He says these 16 hacks “will save you hours of time, grow your business, and boost your productivity in ways that will blow your freaking mind,” and frames Claude as a tool that lets one operator “do 10 times, 100 times more work than anybody else.” Two CTAs are seeded in the middle and end: DM him “AI business” on Instagram to get his AI Company OS Playbook, and click through to his own platform, Apex, marketed as a more secure digital-twin alternative to using Claude Code directly.

His specific claims are concrete enough to check. He tells viewers to import their ChatGPT memory in under two minutes, plug Gmail and Google Calendar into Claude through built-in connectors, schedule recurring Cowork jobs that take over the desktop, install the Chrome extension to scrape and click for them, and use Claude Design (which he calls the newest release) to produce pitch decks and landing pages. He waves off security questions about the Gmail connector with, “Your security at home is a bazillion times less secure than the team at Claude. So stop asking me if it’s secure.”

That’s the line worth pausing on.

What the method actually requires

Most of the 16 hacks are gated behind a paid plan, and a few of the showpieces are still in beta with limits Anthropic itself publishes.

Start with money. Claude Free gets you the chat box and basic models. Claude Code, Cowork, Projects with collaborators, and the higher usage limits start on the Pro plan at $17/month annually or $20/month monthly, per Anthropic’s pricing page. The Max plan, where heavy users actually live, runs $100/month (5× Pro usage) or $200/month (20× Pro usage). Team and Enterprise tiers price per seat above that. So when Martell tells a viewer to “just schedule it” with Cowork or run Claude Code as their dev team, the unspoken floor is roughly $240 to $2,400 a year — before any API spend if you wire Claude into automations through Telegram, Discord, or iMessage as he suggests in hack #14.

Now the Chrome extension. Martell calls it “non-negotiable.” Anthropic’s own launch post describes Claude in Chrome as a research preview that started with 1,000 Max-plan users in August 2025, expanded to all Max subscribers in late 2025, and only widened to Pro, Team, and Enterprise plans in December. In its adversarial testing on 123 test cases across 29 attack scenarios, Anthropic measured a 23.6% prompt-injection success rate without mitigations and 11.2% with them. That’s not a footnote. That’s the company telling you that on roughly one in nine targeted attacks, a malicious webpage could redirect Claude to do something other than what you asked.

Security researchers have validated the worry. In March 2026, The Hacker News reported on “ShadowPrompt,” a zero-click prompt-injection chain in the Claude Chrome extension that combined an over-permissive origin allowlist with a DOM-based XSS in a CAPTCHA component. SecurityWeek covered a separate flaw — dubbed “ClaudeBleed” — that let any other Chrome extension hijack Claude with no special permissions. Anthropic patched both (versions 1.0.41 and 1.0.70), and Claude Opus 4.5 reportedly drives injection success rates closer to 1% in internal testing. But these are not theoretical risks the company hides. Anthropic’s own safety guidance for the Chrome extension explicitly tells users not to use it for “financial, legal, medical, or other types of sensitive information,” and blocks high-risk categories outright.

Re-read Martell’s line about Gmail security in that light. Is your home setup less hardened than Anthropic’s? Probably — CNBC reported Anthropic closed a $30 billion funding round in February 2026 at a $380 billion valuation, so the company has real resources behind its security team. Does that make it fine to point a beta agent at your inbox? Not according to the people who built it.

Cowork and Claude’s computer-use mode carry the same caveat. Anthropic’s documentation recommends running it in “a dedicated virtual machine or container with minimal privileges” and asking a human to confirm consequential actions. The model “might make mistakes or hallucinate when outputting specific coordinates,” in Anthropic’s words. None of that surfaces in the 16-minute tour.

Here’s the rough cost-and-status map a careful viewer needs:

Does the productivity math survive a calculator?

Sometimes. The headline claim that one operator using Claude can do “10× or 100× more work” is the kind of statement the FTC has specifically warned against in its “Keep your AI claims in check” guidance, which tells marketers not to imply performance benefits without “competent and reliable evidence.” That guidance applies to creators and brands alike under the FTC’s endorsement rules — and similar substantiation standards exist for U.K. readers under the ASA, for Australians under ACCC, and for Indian audiences via consumer-protection rules.

What does the real research show? BLS’s productivity research program tracks AI’s measured effect on U.S. labor productivity, and the data so far points to single-digit-percentage gains, concentrated in specific occupations like software development and customer support. Fortune reported in May 2026 on what economists are calling a new “Solow paradox”: labor productivity is up roughly 2.8% over the last year, but total factor productivity has barely moved since the post-pandemic surge. Controlled studies — like the often-cited Harvard Business School consulting experiment — show task-level speedups around 25%, not 1,000%.

Translation: Claude can absolutely shave hours off a week of inbox triage, code scaffolding, and content drafting. It will not turn one founder into a 100-person team. The video conflates the two.

Who actually wins this game

The people getting outsized returns from Claude right now share a profile. They already know what good output looks like in their domain, so they can catch the model when it’s wrong. They write detailed prompts and reusable system instructions instead of one-shotting. They run Claude inside a workflow they understand (a Notion knowledge base, a customer-research pipeline, a coding repo), not as a magic box. And, importantly, they have either the Max plan or business-tier API spend to actually use Cowork, Claude Code, and the Chrome extension at the cadence the video implies.

Martell himself fits this profile — a founder running multiple companies, with a chief of staff, a paid product team, and an existing audience he can convert into Apex signups. That last part is the real funnel.

Is the video useful anyway? Sure, as a feature inventory. As a productivity prediction for a beginner with a Free plan and no audience? Less so.

What you’d realistically earn (or save)

Strip out the “10×” framing and the realistic savings for a small business owner who actively adopts five or six of these hacks looks something like this: 3 to 8 hours per week back from inbox triage, meeting summaries, first-draft writing, and one-off scripts, after a 10 to 20 hour learning ramp. That’s real money — at a $75/hour effective rate, it’s $1,000 to $2,400 a month in time recovered — but it lands well short of “replace your team.”

To capture it, plan on $20 to $200/month for Claude Pro or Max, plus possibly $20 to $200/month in API spend if you wire Claude into Telegram, iMessage, Discord, or scheduled jobs. Compare that to the video’s framing, which treats every feature as essentially free.

Who this is (and isn’t) for

If you already pay for ChatGPT Plus or Claude Pro and you’re a one-person business or a small operator, watching the video and trying three hacks this week is a reasonable use of an afternoon. Memory import, Projects, and Skills will pay back fast. The Gmail and Calendar connectors are fine for a personal account, with the caveat Anthropic itself notes.

If you’re a beginner on the free plan, a creator without a clear workflow, or anyone tempted to point Claude at a work inbox containing client data, the video is selling you the photo and not the recipe. Don’t run computer-use mode on the same machine you do your banking from. Don’t connect Claude to a Gmail account that contains regulated data. And don’t expect Apex, the AI Company OS Playbook, or any other downstream CTA to magically turn the 16 hacks into a business.

What to remember

The features Martell demos are real and mostly worth learning. The framing is the half-truth: paid tiers, beta limits, and ongoing security advisories all get filed under “stop asking, just use the tools.” For more context on Claude’s plan structure and what each tier actually unlocks, see our breakdown of every level of Claude explained. And if you’re weighing this against other “use AI to scale a one-person business” pitches, our piece on how to start a 1-person business with Claude in 30 days goes deeper on the workflow side.