Thehired.ai

My name is Arman the CEO and founder of TheHired.AI based in Dubai.

Chief Executive Officer

Call 15000 leads without calling using AI(Dubai real estate Agencies) We Help Dubai Real estate Agencies Close 20x more clients using AI. Learn How👇👇

We build AI Client Acquisition systems Exclusively for +50M/Year High Ticket Services/product businesses That get them 35% more Qualified meetings.

Arman Liaghat

Exhausted dad of three | Lucky husband to one | Everything else is AI

AI isn't smart, it's fast. What we think of as a human-analogue machine is really just a very advanced probability engine. Without some guidelines to our interaction, it can easily be wrong or just straight up lie to us. And that's the essence of prompt engineering: finding the right phrasing to maximize the probability of an accurate output.

Generative AI doesn't have to be an overly complicated addition to our lives. In fact, in my years of using it, I've seen the conversation around it shift from the technology and lean heavily into the social, cultural, and philosophical aspects.

So if you're a normal human like me who doesn't speak code, follow me for concise, simplified AI news and practical skills that'll keep you afloat on The Coming Wave.

Arturo Ferreira

Someone just posted their ChatGPT news aggregation prompt.
It pulls headlines from 5 sources in one search.
Reuters, Forbes, Fox News, Yahoo, Bing.
Then synthesizes the top 3 topics across 4 categories.

Most teams ask ChatGPT one question at a time.
This person built a systematic workflow.
That's the difference between a tool and a system.

Here's how to build your own research workflow:

1. Map your current research process
List every source you check manually each week.
Industry news sites, competitor blogs, trade publications, regulatory updates.
Track how many hours your team spends aggregating this information.
Most teams waste 5-10 hours per week on manual research.

2. Build a sequential prompt structure
Start with "Follow these steps in order."
Add numbered tasks for each source to check.
Include "DO NOT SHOW ME YET" after each step.
This forces ChatGPT to complete all research before summarizing.

3. Add exclusion logic to prevent duplicates
Tell ChatGPT to skip sources already covered.
Example: "Exclude headlines from X, Y, Z since we covered those."
This prevents repetitive information across sources.
Saves 30-40% of output length.

4. Define your output format upfront
Specify categories: competitors, regulations, market trends, technology.
Request specific structures: newsletter, brief, comparison table.
Set priority levels: critical, monitor, background.
ChatGPT defaults to generic summaries without this.

5. Test which sites allow ChatGPT access
Not all websites permit web browsing tools.
Run test searches on your target sources.
Build a vetted list of accessible sites.
Rotate blocked sources with industry alternatives.

6. Schedule this as a recurring workflow
Run the same prompt weekly or daily.
Save approved prompts as templates.
Assign team members to review specific categories.
Your research becomes systematic, not ad hoc.

A junior analyst doing this manually costs $50K/year.
This prompt costs $0.20 in API fees per run.
Most teams are still paying humans to copy-paste headlines.

The people who win with AI aren't finding better tools.
They're building better systems with the tools everyone has.

Found this helpful? Follow Arturo Ferreira.

How to turn ChatGPT into a $10K/month research analyst

Most analysts spend 3 hours in Excel.

Building pivot tables. Creating charts. Debugging formulas.

You open the CSV file.
Filter columns manually.
Build calculations one by one.
Copy charts into PowerPoint.

AI agents do this entire workflow autonomously.

Goose is an open-source agent that handles end-to-end analysis.

The 4-step workflow that eliminates manual data work:

1. Ask for analysis in plain English

Drop your CSV file in the directory.

Prompt: "Tell me the best revenue items and busiest days. What are the trends?"

That's it. No SQL. No Python. No formulas.

2. Watch Goose orchestrate everything

It uses ripgrep to locate your file.
Generates a Python script with Pandas.
Creates and debugs virtual environments automatically.

You don't write a single line of code.

3. Review the generated insights

Goose delivers text-based findings.
Answers your specific questions.
Adds unsolicited recommendations from patterns it found.

4. Generate shareable HTML dashboards

Follow up: "Turn this into a webpage with interactive charts."

Goose uses Plotly to visualize data.
Creates a professional HTML report.
Opens it in your browser automatically.

Your analysts are building infrastructure when they should be finding insights.

Every pivot table is 20 minutes gone.
Every chart reformatted is time not spent thinking.

Goose handles the technical overhead autonomously.
Your team focuses on what the numbers mean, not how to calculate them.

Found this helpful? Follow Arturo Ferreira

How to Turn CSV Files Into Reports in 5 Minutes

Most executives wing their stakeholder meetings.

Then wonder why nothing gets approved.

You walk into a boardroom.
Three executives with competing interests.
You have 30 minutes to get buy-in.

You're making it up as you go.

Elite communicators built something different:

An AI coach trained on Dale Carnegie and Jane Austen.

The 5-step system that turns you into a stakeholder whisperer:

1. Craft a master prompt for influence

Use Anthropic Console's Prompt Generator.
Create an "Influence and Communication Coach" persona.

Define the inputs: situation, stakeholders, goals.
Define outputs: analysis, strategy, play-by-play.

This becomes your strategic advisor.

2. Train it with classic persuasion literature

Upload public domain books to a Claude Project.
"How to Win Friends and Influence People" by Carnegie.
Jane Austen novels for navigating social dynamics.

Your AI now has 200 years of persuasion expertise.

3. Simulate your actual scenario

Describe your high-stakes meeting.
List the stakeholders and their motivations.

Example: "Product launch delayed. VP of Sales wants to push anyway. VP of Engineering needs 2 more weeks. CEO presenting to client in 3 days."

4. Generate a strategic playbook

The AI creates a comprehensive plan.
Situation analysis with psychological insights.
Multiple strategic approaches ranked by risk.
Minute-by-minute meeting agenda.

You get a roadmap, not generic advice.

5. Prepare for curveball questions

Ask: "What unexpected questions might come up?"

The AI anticipates objections.
Suggests responses that redirect without deflecting.

You walk in with answers to questions that haven't been asked yet.

Here's what changes:

You stop reacting to stakeholder dynamics.
You start orchestrating them.

The VP of Sales feels heard before pushing back.
The VP of Engineering gets credit for the compromise.
The CEO sees you as the person who makes problems disappear.

Why classic literature matters:

Carnegie wrote the playbook on influence in 1936.
Austen understood stakeholder politics in 1813.

These principles haven't changed.
Your AI coach applies them to your specific situation in real-time.

You get centuries of wisdom customized to your 3pm meeting.

Most executives think stakeholder management is a soft skill you either have or don't.
When it's actually a learnable system that AI can coach you through.

Found this helpful? Follow Arturo Ferreira

How to Prepare for High-Stakes Meetings in 15 Minutes

Your hiring decisions are inconsistent.

Because your interview feedback is all over the place.

One interviewer writes novels.
Another writes three bullet points.
Nobody follows the rubric.

Your hiring committee can't compare candidates fairly.

Here's what high-growth companies built:

A custom GPT that standardizes feedback automatically.

The 4-step system that makes every scorecard useful:

1. Build a detailed rubric first

Define the competencies you're measuring.
Outline what good looks like at each level.

This becomes your source of truth.
Not just another document nobody reads.

2. Create the GPT prompt with context

Feed it your complete rubric.
Include examples of excellent scorecards.
Include examples of terrible scorecards.

Specify the output format you want.
The AI learns from good and bad patterns.

3. Run scorecards through the GPT

After each interview, the interviewer submits their notes.

The GPT analyzes against the rubric.
Rates the feedback quality.
Identifies what's missing.

4. Get standardized output automatically

The GPT generates structured feedback.
Highlights candidate strengths and gaps.
Creates a Slack-ready summary.

Every scorecard now follows the same format.
Your hiring committee can actually compare candidates.

The hidden cost of inconsistent feedback:

You hire based on who wrote the best scorecard.
Not who was the best candidate.

Verbose interviewers sound more confident.
Laconic interviewers get ignored.

Neither style tells you if the candidate can do the job.

AI enforces the rubric when humans forget to.
It catches missing competency assessments.
It pushes interviewers to be specific instead of vague.

You're competing for talent against companies with consistent hiring processes.
While your feedback quality depends on who's in a chatty mood.

Found this helpful? Follow Arturo Ferreira

How to Fix Inconsistent Interview Feedback

MIT just published research on why ChatGPT struggles with state tracking.
The problem isn't memory.
It's how transformers encode position information.

Current models use RoPE (rotary position encoding).
It treats all words four positions apart the same way.
Doesn't matter if it's "cat sat on box" or financial data changing over time.

MIT-IBM Watson AI Lab built PaTH Attention to fix this.
It outperforms RoPE on state tracking and sequential reasoning.

Here's what this means for how you use LLMs today:

1. Audit where your LLM loses context in long documents
Test with financial reports, legal contracts, or multi-step instructions.
Track where the model misses state changes or sequential logic.
Example: "Company X acquired Y in Q2, then sold Z in Q4" often gets confused.
Current position encoding can't track entity relationships over time.

2. Break complex documents into state-aware chunks
Don't feed 50-page contracts as single prompts.
Segment by state changes: before acquisition, during transition, after close.
Explicitly label each section's timeframe and context.
This compensates for positional encoding limitations.

3. Use explicit state markers in your prompts
Add "Current state:" before each major transition.
Example: "Current state: Post-merger. Previous state: Pre-merger."
Forces the model to treat position changes as data, not just distance.
Reduces errors in multi-step reasoning by 40-60%.

4. Test LLM performance on conditional logic tasks
Build test cases with "if-then" sequences over long contexts.
Example: "If condition A occurs on page 5, apply rule B on page 20."
Current models fail these because RoPE doesn't track causal relationships.
Know your model's limits before deploying in production.

5. Prioritize reasoning over retrieval for complex documents
RAG (retrieval-augmented generation) won't fix state tracking issues.
It retrieves chunks but doesn't understand how states evolve.
For contracts, regulations, or multi-step workflows, use specialized parsing.
Position encoding is the bottleneck, not retrieval accuracy.

6. Watch for next-gen models with adaptive position encoding
PaTH Attention is research, not yet in production models.
But it signals where LLM architecture is heading.
Models that track state changes will replace current transformers.
Plan your document processing stack accordingly.

Why this matters:
You're using LLMs on tasks they structurally can't handle well.
Financial analysis, legal review, code debugging over long contexts.
All require state tracking that RoPE fundamentally doesn't provide.

MIT just showed the problem and the solution.
Most teams won't adjust their workflows until new models ship.
You can compensate for these limitations now.

Found this helpful? Follow Arturo Ferreira.

Why your LLM keeps losing track of critical details in lon

You're in 47 WhatsApp groups.

3,000 unread messages. Zero chance of catching up.

Your team coordination happens in group chats.
Client updates arrive via WhatsApp.
Family planning runs through message threads.

You've given up on reading most of it.

Technical founders built something different:

They connected WhatsApp directly to Claude using MCP.

The 5-step system that turns message chaos into searchable intelligence:

1. Clone the open-source WhatsApp MCP project

Download Luke Harries' whatsapp-mcp from GitHub.
Run the script locally in your terminal.

It uses the whatsmeow library to simulate WhatsApp Web.

2. Sync your message history

Scan the QR code with your phone.
Just like connecting WhatsApp Web.

The script performs a one-time sync.
Stores everything in a local SQLite database.

Your messages never leave your machine.

3. Run the local MCP server

The script creates a secure bridge.
Exposes functions like getRecentMessages and sendMessage.

Claude can now interact with your WhatsApp data.
Without your messages hitting external servers.

4. Connect to Claude

Link your local MCP server to Claude.
The AI gains access to the exposed functions.

It can read your chats.
It can send messages on your behalf.

5. Query with natural language

Ask: "Summarize my work group chat from this week."

Or: "What did the design team decide about the rebrand?"

Claude scans your messages and extracts the answer.

What this changes:

Group chats with 500 unread messages become searchable databases.
Client communication gets automatically categorized.
Project decisions buried in threads surface instantly.

You stop scrolling through noise.
You start asking questions and getting answers.

Found this helpful? Follow Arturo Ferreira

taplio

The Battle

Arman Liaghat

Arturo Ferreira

Followers

Posts

Engagements

Likes

Performance ScoreArturo Ferreira's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Arman Liaghat

Arturo Ferreira

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Arturo Ferreira's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreArturo Ferreira's LinkedIn Performance