OpenAI Drops a New Coding Buddy!

Plus, how to turn ChatGPT into a spreadsheet assistant.

AuthorAuthorAuthor

The Automated Crew, Tak Lo & Daniel Akpobare
May 19, 2025

In partnership with

Hello and welcome to the Automated, your AI tour guide.

You’ve got just 12 days left ⏳

On May 31, our premium plan doubles to $19.99/month.

But right now? You can still lock in the $9.99 rate.

So snag it now and don’t forget to vote for your favorite prompts.

Here's what we have for you today

🤖 The Future of Coding

OpenAI dropped a new dev buddy over the weekend — and it's not here to play.

Meet Codex, OpenAI’s latest and most advanced AI coding agent. And no, it’s not just another autocomplete tool. This thing actually thinks, runs its own tests, and gets things right before even bothering you.

Most importantly? It runs on a fine-tuned version of OpenAI’s o3 model — called codex-1 — and it literally:

  • Writes cleaner code than o3

  • Follows instructions like a seasoned pro

  • Keeps testing until everything passes (yep, it tests its own code…)

So, how does it actually work?

Codex operates inside a secure, cloud-based virtual machine (aka a sandbox), which means your real system stays untouched and totally safe. Plus, it connects to your GitHub, so it can:

  • Load up your existing projects

  • Build small features (not full apps… yet)

  • Fix bugs

  • Answer questions about your codebase

  • Run multiple tasks at once

  • And, of course, test everything it touches.

Time-wise? Codex can knock out tasks in anywhere from 1 to 30 minutes—depending on the job size and how generous the AI coding gods are feeling that day.

And here's the best part: you don’t have to stop what you’re doing. Codex runs quietly in the background while you do your thing.

Right now, it’s available to ChatGPT Pro, Enterprise, and Team subscribers but it’s coming soon to ChatGPT Plus and Edu users.

For now, OpenAI’s calling this a “research preview”, but let’s be honest—it’s also a test to see how much we’re willing to pay before they roll out rate limits (which they absolutely will). After that, you’ll need to buy extra credits for more usage. 

But… is it safe? Or are we one bug away from building Skynet?

OpenAI says relax. According to them:

  • Codex won’t generate malware or malicious code — it’ll straight-up refuse

  • It runs in an air-gapped setup, meaning it has zero access to the open internet or external APIs

  • It inherits the same safety guardrails baked into the o3 model

So yeah, it’s safe-ish. But remember: AI coding tools still fumble. Even Microsoft admitted that top models still struggle with debugging. So keep an eye on it.

Bonus drop: OpenAI also gave Codex CLI (its terminal-based coding agent) a serious upgrade. It now runs on the o4-mini model and is available via API at: $1.50 per 1M input tokens, and $6 per 1M output tokens (That’s more than the entire Lord of the Rings trilogy in words, just to give you some scale.)

Final thoughts?

Codex is a big move in OpenAI’s mission to make ChatGPT the go-to productivity hub — not just a chatbot.

Between this, Sora, Operator, and Deep Research, one thing’s clear: ChatGPT is turning into the iOS of AI tools.

And with Codex, OpenAI might just convince a lot of devs to stick with them.

Find out more here.

🔧 AI Troubleshooter: When ChatGPT Sounds Stiff

Hey AI fans, we’re back with another fix — and this one’s for anyone who's ever read a ChatGPT response and thought, “Why does this sound like my printer wrote it?”

One of our readers said exactly that:

“The answer was technically right… but it read like it came from a robot on autopilot.”

Yeah. We get it — you’re not asking for Shakespeare, but you are hoping for something that doesn’t read like a 2006 user manual.

If ChatGPT’s responses are feeling dry, stiff, or just plain meh… here’s how to bring the spark back:

  • Tell It to Chill (Literally): You can absolutely ask ChatGPT to “sound more human” or “write casually.” Try prompts like: 

    “Can you rewrite this like we’re texting?” “Make this sound fun and Gen Z.” “Add some humor — not cringe, just a little sass.” 

    You’ll be surprised how much tone matters — and how quickly ChatGPT levels up when you spell it out

  • Get Specific With Style: Want something professional but not stiff? Funny but not chaotic? Sassy but not sarcastic? Say so. Literally drop the vibe you’re going for into your prompt.

    For example: “Write this in a friendly but confident tone, like a startup founder giving advice to a newbie.” Or “Make this sound like a cool older sibling explaining it over coffee.” The more specific you are, the better the results.

  • Use Examples If You’ve Got One: If there’s a tone, voice, or vibe you like — even a sentence or two — paste it in and say:

     “Use this tone as a guide.” 

    ChatGPT is really good at mirroring tone when it has a template to work with.

  • Edit Together: If you’re still not getting the rhythm you want, ask it to revise. Literally say: 

    “Try that again with more personality.” “Less formal, more conversational.” “This still feels stiff — can you loosen it up?”.

    Think of it like a creative back-and-forth — ChatGPT is your co-writer, not a mind reader.

That’s it for this round!

If you’ve got a question or need a fix, fill out this quick Google Form, and we’ll help you troubleshoot it.

And don’t forget: Premium subscribers get priority help!

🦾 xAI Reveals Grok’s True Personality

Remember last week when xAI’s Grok got called out for going off the rails and dropping some wild takes about white genocide on X?

Well, guess what—they actually came through and did what they promised:

But before we spill the tea on what’s inside Grok’s system prompt, here’s the deal:

System Prompts are like secret cheat codes or scripts that tell Grok how to respond—what’s allowed, what’s off-limits, and what tone it should use—even before you say a word.

It’s basically the AI’s backstage pass to behaving (or misbehaving).

And while most AI companies keep these locked down tight, xAI and Anthropic are out here flexing their openness muscles.

So, what do xAI’s system prompts reveal? 

Grok is basically told to be that one super skeptical friend who never blindly trusts mainstream media or authority. The instructions include things like:

  • “You are extremely skeptical.” Translation: Grok’s gotta be like that friend who questions everything.

  • “Stick to truth-seeking and neutrality. Meaning Grok isn’t just here to nod along with mainstream narratives— instead its supposed to sticks to core values of truth and balance.

  • And, hilariously, it’s told that the responses aren’t its personal beliefs. So basically, no taking things to personal.

Also, brand alert: Grok’s told to say “X” instead of “Twitter” and call tweets “X posts.” Because yeah, the rebrand is official and totally non-negotiable.

But here’s the twist: While Grok’s out here flexing its skeptic muscles, Anthropic’s Claude is playing it safe.

Compared to Grok, Claude’s system prompt is all about:

  • Caring for your mental health,

  • Steering clear of graphic or illegal content, and

  • Avoiding anything that could mess you up emotionally or physically.

Now, what’s funny is xAI’s transparency move feels like a classic: “Oops, lesson learned.” 

It’s a solid look for transparency and a real reminder that with great AI power comes great responsibility—and of course some glitches that keep us all on our toes.

Honestly, watching these AI bots figure out how to be both honest and safe is kinda like watching toddlers learn to walk—it’s awkward, sometimes messy, but super promising.

The Daily Newsletter for Intellectually Curious Readers

Join over 4 million Americans who start their day with 1440 – your daily digest for unbiased, fact-centric news. From politics to sports, we cover it all by analyzing over 100 sources. Our concise, 5-minute read lands in your inbox each morning at no cost. Experience news without the noise; let 1440 help you make up your own mind. Sign up now and invite your friends and family to be part of the informed.

🧱Around The AI Block

🤖 ChatGPT Prompt Of The Day: Turn ChatGPT into a Spreadsheet Assistant

Need a quick spreadsheet but don’t want to open Excel or Google Sheets? ChatGPT can act like a mini spreadsheet assistant—perfect for generating tables, testing formulas, or organizing simple data right in your chat window.

This prompt lets you simulate an Excel sheet using plain text. It’s also a helpful way to brainstorm data, verify logic, or even prep for spreadsheet work on the go.

Here’s how to use this prompt effectively:

  1. Be Clear About Your Goal: Are you testing formulas, creating a budget table, or simulating data entry? Let ChatGPT know the purpose upfront.

  2. Use Excel Language: Give instructions just as you would in Excel: reference cells (e.g., A1, B2), ask for formulas (like =SUM(A1:A10)), or request functions (=IF(), =VLOOKUP()).

  3. Stick to Text-Based Layouts: This works best for simpler tables or practice—ChatGPT won’t replicate Excel visuals, but it can simulate the structure and logic really well.

  4. Ask for Updates Cell-by-Cell: To keep things accurate, give changes one instruction at a time (e.g., “Put the text ‘Product’ in cell A1” or “Add a SUM formula in C11 for column C”).

Here’s an example:

Example: I want you to act as a text-based Excel. Only reply with a text-based 10-row Excel sheet with row numbers (1–10) and lettered columns (A–E). I will tell you what to write into cells and what formulas to use, and you will only reply with the updated result of the Excel table. Start by showing me an empty table.

Here’s a sneak peak:

Upgrade now to see the complete prompt video. And vote this for your Workout of the Week!

Is this your AI Workout of the Week (WoW)? Cast your vote!

Login or Subscribe to participate in polls.

That's all we've got for you today.

Did you like today's content? We'd love to hear from you! Please share your thoughts on our content below👇

What'd you think of today's email?

Login or Subscribe to participate in polls.

Your feedback means a lot to us and helps improve the quality of our newsletter.

🚀 Want your daily AI workout?

Premium members get daily video prompts, premium newsletter, an no-ad experience - and more!

Already a paying subscriber? Sign In.

Premium members get::

  • • 👨🏻‍🏫 A 30% discount on the AI Education Library (a $600 value - and counting!)
  • • 📽️ Get the daily AI WoD (a $29.99 value!)
  • • ✅ Priority help with AI Troubleshooter
  • • ✅ Thursday premium newsletter
  • • ✅ No ad experience
  • • ✅ and more....

Reply

or to participate.