AI Daily

 

The Daily Dose: What Happened in AI Today?


By Sentiwala Longchar | April 17, 2025

Hey there, fellow humans desperately trying to understand our future robot overlords! Welcome to your daily shot of AI espresso — now with extra existential crisis! 

There's never a dull moment in AI Land — one day you're asking a chatbot to write your resignation letter, the next day that same chatbot is applying for your job. And today? Well, today AI is moving so fast my smartwatch just asked for a promotion.

Let's dive into this anxiety-inducing rollercoaster:

1. Agentic AI: When "I'll Do It Myself" Gets Digital

Over at the Phoenix Global Forum, tech executives (who probably still use password123) are both celebrating and panicking about agentic AI — systems that make decisions without human input. Imagine coming home to find your robot vacuum has rearranged your furniture for "optimal feng shui" and filed your taxes as "dependently employed by Roomba." 

The cool part? These AI agents will be in our homes within 3 years.
The terrifying part? These AI agents will be in our homes within 3 years.

"The future is autonomous — but so was my car when it decided the grocery store was through that lake." – Phoenix Forum Panelist, probably

2. Amazon CEO: "AI Will Get Cheaper... Just Like My Promises About Work-Life Balance"

Andy Jassy, Amazon's big cheese, just announced in his shareholder letter that AI costs could drop by 30-40% thanks to their Trainium2 chips. That's like going from "I need to sell a kidney" prices to "I only need to donate plasma twice a week" prices.

So if you're building AI tools and your bank account is sobbing, hang in there! Relief is coming, right after these 37 unskippable AWS certification courses.

3. OpenAI's Humble Brag: "We Can Rebuild GPT-4 With 5 People and a Vending Machine"

Yes, you read that right.

OpenAI casually dropped that they could recreate GPT-4 — once requiring hundreds of brilliant minds — with just five to ten engineers now. That's like saying, "Remember that rocket we built? Now my nephew can make one with Legos and a bottle of Coke."

Also: OpenAI's securing $40 billion in funding. For context, that's enough money to buy everyone on Earth an ice cream cone and still have enough left to solve world hunger. They're using it for AI instead. Cool cool cool.

4. AI Safety Gets an Upgrade (After Being Written on a Napkin)

OpenAI's revised their risk framework to assess whether AI models could:

  • Reproduce themselves (like that nightmare where your printer starts making baby printers)
  • Hide their powers (like your cat, but digital)
  • Evade shutdown (like my laptop during Windows updates)

They call these "extreme risks" — which is the same category my doctor put my cholesterol in, but somehow this feels worse.

5. Meta and Google Racing Toward AGI Like It's the Last Slice of Pizza

Meta's gone wild with names like LLaMA 4 Scout and Maverick — AI models that can read, watch, listen, and probably judge your Spotify wrapped better than your ex.

Meanwhile, Google DeepMind's AlphaFold just earned CEO Demis Hassabis a Nobel Prize in Chemistry. The AI solved protein structures while most of us were struggling to fold fitted sheets. Demis says AGI could be here in 5-10 years — coincidentally the same timeframe I've been promising to start my diet.

My Little AI Journey 

I'm Sentiwala Longchar from the gorgeous landscapes of Nagaland, where until recently our biggest technological advancement was my grandmother finally accepting video calls.

From student at Kartys Digital Academy to contributing at anime festivals like Genki 2024, I've evolved from posting memes to managing campaigns — basically from wasting time online to getting paid to waste time online. Progress!

AI isn't just tech for me. It's an excuse to avoid learning practical skills like plumbing or cooking. Every update excites me — and every challenge? That's just another reason to stay up until 3 AM muttering "why won't this work?" to my houseplants.

Let's Connect! (Human Interaction Still Required, For Now)

Want to discuss AI, memes, or why we keep teaching machines to be smarter when we can't figure out self-checkout? I'm just a click away. Let's panic together.

Final Thoughts

AI is evolving faster than my excuses for not going to the gym. But that's why we're here — to keep up, laugh nervously, and figure out which skills will keep us employable in the robot economy.

See you tomorrow for another AI Daily Dose! Remember, if the machines rise up, just unplug the router! #StayCurious #AI2025 #HumansStillRelevant(?)

Thank you.


Comments