- H4CKER BY NATHAN BINFORD
- Posts
- THIS WEEK: THE AGENTIC FUTURE IS NOW
THIS WEEK: THE AGENTIC FUTURE IS NOW
Holy moly, Perplexity's Comet browser has opened my eyes. We're doing everything wrong.

🌟 Editor's Note
Thanks for reading H4CKER, my weekly newsletter on AI, marketing, and automation. Please share this with anyone you think might find it interesting and let me know what you think about my hot takes on these timely topics.
Drive leads and revenue on autopilot with my proven growth marketing system. Just Subscribe And Grow: no contracts, no BS, just leads that convert.
☄️ Perplexity’s Comet Changed The Game
Perplexity, the so-called “search wrapper for ChatGPT” that nearly everyone continues to underestimate, started releasing the world’s first agentic browser recently; using an invite system to control the potentially explosive demand.
The world’s already changed but no one knows.
It’s wild to see people just going on with their lives, using Chrome to do repetitive manual tasks, not knowing that the future is already here -that your web browser is intelligent and it can do your work for you.
Ok, this may be hyperbole, but only by a little. Not everyone has access to Comet and surely, at some point, Perplexity is going to have to charge more for it, because people (like me) will use it for multi-step workflows with reasoning involved and the costs add up.
Right now, with minimal users, the caps are generous for invited Pro subscribers (estimated around 300 queries / actions per day) and unlimited for Max subscribers.
Pro costs $20/m and Max, $200/m.
But what really matters here is that we’ve crossed the bar. Once you get a taste of what’s possible, you can’t just go back to manually grinding out repetitive tasks in your browser.
You’ve tasted the future. You know what’s possible…
Agentic AI Requires Us To Think (And Work) Differently
We’re doing everything wrong.
We’re hiring and staffing wrong.
We’re researching and planning wrong.
We’re building and testing and deploying wrong.
Wrong. Wrong. Wrong. But it’s all about to change.
$200/m for what the current top-tier AI plans are providing in value is going to feel cheap because we’ll be replacing $1,200 in virtual assistant costs in the same move.
And that’s just the beginning. Comet, and other agentic tools, will rapidly replace a lot of the simpler automations that we need to maintain; or at least incentivize upgrading them with agentic features versus linear automations.
But agentic AI won’t just replace the old way of doing, it’s also changing our way of thinking as well.
Every activity turns into a prompt engineering exercise first and a physical task, second.
Taking full advantage of these tools requires thinking out every activity in advance, describing it effectively to the agent, and providing clear enough instructions to allow the agent to perform well in every scenario.
It’s a little bit like giving instructions to a freelancer half a world away; you get the best results when you’re:
clear
explicit
detailed
giving examples
accounting for edge cases
Working with agents means prompting effectively for tasks you’ve always done through muscle memory; and that can be surprisingly challenging.
But one thing is plainly obvious to anyone that’s paying attention: prompt engineering is most important skill in the world for anyone who’s currently a “knowledge worker”.
What got us here, won’t get us where we’re going.
📣 Incredible Use Cases For Comet, ChatGPT Agent, And Other Agentic AI
Automating research and synthesis across multiple tabs and sources
Real-time summarization of articles, videos, and podcasts
Intelligent shopping: product comparison, price tracking, and purchasing
Email management: triage, priority sorting, drafting responses, and unsubs
Scheduling automation: finding times, sending invites, handling calendars
Social media: creating, and posting content; managing engagement
Automated content creation workflows and competitor analysis
Lead generation and prospect research
Automated recruitment workflows via platforms like LinkedIn
Study assistance for students (summaries, study guides, quiz generation)
Language translation and simplified explanations for immersive learning
Comprehensive travel planning: research, itinerary creation, and booking
Personal task automation: appointments, lists, calendar management
Developer support: technical research, code analysis, API breakdowns
Automated financial and investment research and reporting
I mean, wow. What can you say to that but wow. The potential is staggering, the inevitability terrifying, and size of what’s possible now, is rapidly expanding.
My next 3-4 computer purchases will probably be for agents.
Welcome to our agentic future. LFG! 🚀
⚙️ The Laboratory: Prompts & Automations
This week I took Perplexity Comet on a test drive, documenting three absolutely mind-blowing use cases that will have you rethinking the way you work.
✅ First Task: Searching ExpiredDomains.net for potential PBN domains and using Ahrefs to assess their potential.
✅ Second Task: Examining a spreadsheet of content ideas and using it to brainstorm an additional 5 rows for the same sheet.
✅ Third Task: Drafting a job posting on Upwork.
It’s most remarkable how immediately useful Comet is. It didn’t take much trial and error to generate results that were valid and useful. If you’re halfway decent at prompting you can just download and go.
Find out how AI can help your streamline and scale your business, cut costs, and drive more revenue -schedule a free audit today!
🚀 AI / Marketing News
Google’s Secret Sauce: 1,200 Experiments, AI Agents, and Deep Entity Mapping Revealed

A deep trove of research on Google’s current expansions in search technology have revealed the extent of their obsession with entities, their attempt to catalog the world, and personalize search to each and every person on earth.
Source: SearchEngineLand
Takeaways:
Google runs ~1,200 continuous experiments
Details confirm the 2024 leaks and revealing more
Entities are the backbone of Google via the Knowledge Graph
New systems like Livegraph and ghost entities form real-time responses
AI Mode: 90+ reasoning agents with personalized, session-level context
Discover and other services are tailored to personal behaviors and location
Real-time query expansion dissects intent; scores on context / relevance
My Take:
This inside scoop is the closest we’ll ever get to dissecting the guts of Google Search!
Wondering why search is so volatile these days? Google's not just ‘tweaking the algorithm’; they’re running military-grade A/B testing across the semantic universe and building a matrix of embedded preferences and ambient context on every individual on the planet (connected in any way to Google).
The Knowledge Graph isn’t a helper tooltip anymore; it’s the spine of the whole monstrous organic-AI hybrid called search. And all their widgets, doodads, and twiddlers gaming the SERPs are appendages to this entity-based catalog of everything.
Which means the game now is branded entity validation, vectorized alignment, and baking your site into their AIs’ understanding of the web.
If Google doesn't 'recognize' you, you might as well be invisible. Our job as marketers is to hack this matrix, and assert our own controls so that Google doesn’t get to define reality completely.
Consumers Are Letting AI Pick Their Brands And Double-Checking It

Yext did a study recently on the changing ways that people search the web and discovered that AI is being used by more and more consumers and that “search” isn’t just in Google anymore.
And this isn’t doesn’t seem to be confined to “the bubble” on X, whatever naysaysers may say.
Source: Yext
Takeaways:
62% of consumers trust AI as much as traditional search
43% use AI tools daily; 75% increased AI use compared to last year
68% use AI for local searches; only but 19% trust the results
Millennials use both: AI for quick hits, search for details (69%)
Gen Z favors AI for brainstorming/how-tos and social media for ideas
Gen X leans on AI for summaries, search for general info (70%)
Boomers are mostly holdouts: 80% rely on search, 26% avoid AI
10% of searchers blindly trust AI; 48% verify on multiple platforms.
My Take:
AI’s impact on customer journeys is significant, with a majority of consumers now trusting AI as much as Google.
Nearly half of users verify answers in multiple places, but trust in AI still drops off when the queries are about local businesses.
The generational breakdown is pure gold for marketers:
boomers are still playing safe with search
Gen Z mixes AI with TikTok vibes
millennials do a little of both
If you're not structuring your content for both AI and traditional search, you're basically invisible to one or both.
TL;DR: AI’s becoming the co-pilot of discovery, but no one's just handing the controls over. Make your brand accessible to AI, trustworthy, and consistent across every channel.
AI Models Are Smarter in 2025, But Deployment Anxiety is Higher Than Ever

Iconiq Capital’s 2025 State of AI report, The Builder’s Playbook, takes a look at how companies are adopting AI in their production workflows and where trust has not yet been earned.
Source: Iconiq Capital
Takeaways:
Current LLMs have gotten much better at avoiding hallucinations
Still, 39% of surveyed companies cite hallucinations as their #1 concern
The core issue isn’t technical…it’s trust.
The higher the stakes, the lower the tolerance for AI error.
Companies that succeed prioritize :
domain-specific training
building human-in-the-loop workflows
leveraging confidence scoring
rolling out gradually
Explainability / trust concerns are highest in legal, health, and finance
AI-related engineering will reach 37% by 2026
Average companies use 2.8 models per task.
The best performers follow a 3-layer AI architecture:
customized model selection/training
robust system design
UX built around feedback
Most hallucination issues stem from using overgeneralized models
Training (and providing data) are the real fix
My Take:
Companies are still terrified of AI's hallucinations despite models getting way better; but (surprise!) most of that fear is self-inflicted by lazy training, a lack of data, and relying on generic models.
If hallucinations are a showstopper for you, as a builder, you’re NGMI.
But what’s more interesting in this report is the focus on agentic workflows, not only in AI-native companies (building AI tools) but in the AI-enabled ones that represent the mid and enterprise market.
Agentic AI isn’t just a buzzword, it’s the next big thing.
Google AI Overviews Slash Organic CTRs by 32%

The data is in. AI Overviews have slashed organic search traffic by roughly half. The zero-click era is here. We cannot expect to get traffic from search in the same way we did before (and maybe at all).
Source: Search Engine Journal
Takeaways:
Google's top organic click-through rate (CTR) dropped over 32%.
This coincides with the rollout of AI Overviews
AI Overviews push organic links further down the page
Marketers and publishers are both seeing huge drops in traffic
My Take:
AI Overviews pull the rug out from under SEO strategies.
A 32% haircut off the top spot’s CTR isn’t a glitch, it’s a strategic assault on organic search as a marketing channel.
If you've been living off search traffic, it's time to diversify your lead gen game or start paying the toll to play in Google's walled AI garden.
The AI war for attention has officially hit search, and Google’s playing for keeps.
White House Unveils Aggressive ‘AI Action Plan’

The White House released it’s guidance on artificial intelligence development in the US, with a less is more / hands-off strategy that fully supports AI and prioritizes it as a matter of national security in a new tech cold war with China.
Source: ArtificialIntelligence-News.com
Takeaways:
The AI Action Plan frames AI development as a national security priority
Three core pillars:
accelerate domestic innovation
build national AI infrastructure
assert global leadership
Regulatory rollback planned to empower private sector innovation
Federal funding cuts can be used to block state-level AI laws
Build, Baby, Build: Big investments in data centers, chips, nuclear
Exporting U.S. AI to allies and countering China are key pillars
Risks acknowledged, but less regulation preferred
OpenAI’s Sam Altman and watchdogs express unease
My Take:
The U.S. government just launched into full digital manifest destiny mode.
The new plan throws out red tape, loudly backs the private sector, and basically says, “Get on board, or get out of the way!"
It's an infrastructure-and-export power play crossed with techno-nationalism; complete with energy grid overhauls and a NatSec revamp.
Get in losers, we’re building Stargate!
Practically speaking, this means money will flood into AI infrastructure and public adoption will skyrocket.
There will be a tsunami wave of AI-spurred disruption from the eruption of government support (and great PR) for Silicon Valley.
Will you be ready to ride the wave?

Nathan Binford
AI & Marketing Strategist
I hope you enjoyed this newsletter. Please, tell me what you like and what you don’t, and how to make this newsletter more valuable to you.
And if you need help with AI, marketing, or automation, grab time on my calendar for a quick chat and I’ll do my best to help!