Leadership CONNECT: 27-MAR-2025 (Y25W14)

Greetings AI Thinkers,

This week’s Leadership CONNECT:

  1. Spark of  the week: AI + Human Work = “Vibing” (Source: Yesha on Human Thinking)
  2. Video: Vibe Coding Is the Future (Source: Y Combinator, YouTube, 31m)
  3. New Model: Google Gemini 2.5 (Source: Yossi Matias, LinkedIn, 5m Read)
  4. New Image Tool: Open Introducing 4O Image Generation (Source: OpenAI, YouTube, 16m)
  5. MindLi Update: Feed Your AI FOMO With Our LinkedIn MindLi AI Group (Source: LinkedIn)

Happy Thinking,
Dr. Yesha Sivan and the MindLi Team

P.S. Comments, ideas, feedback? Send me an e-mail.

1. Spark of the Week: AI + Human Work = “Vibing” (Source: Yesha on Human Thinking)

Source: A person sitting at a work desk with DJ tools, he is vibing. Realistic. Orange colors. (MindLi)

Opening: To Vibe, to Be In the Flow, in the Zone

I used Google and OpenAI to generate this image in under 2 minutes. I uploaded a photo and added the prompt: “A person sitting at a work desk with DJ tools, he is vibing. Realistic. Orange colors.” Then, using OpenAI, I layered the text on the shirt: “Yesha is Vibing.”

Source: Using OpenAI, I layered the text on the shirt: “Yesha is Vibing.” (MindLi)

This is a typical example of vibing in the context of image generation — fast, creative, and collaborative.

Andrej Karpathy (AI expert, OpenAI co-founder, ex-Tesla) once tweeted something that captures the essence of vibing in the context of programming:

“The best feeling in programming is when you’re just vibing — code is flowing, everything is working, and you’re in the zone.”

Programming is a standout example of a domain transformed by AI: its structured logic is ideal for AI to analyze, generate, and enhance code. Tasks like refactoring, documentation, and debugging are now faster and more reliable. With AI, programming becomes less about syntax and more about design and thinking. (I’m closely watching this shift and will share more soon.)

One Example: Voice Brainstorming With OpenAI

Yesha said:

“Okay, so there’s a new term that people are now talking about called vibing. Can you define it in two sentences?”

ChatGPT said:

Sure! “Vibing” refers to a state of relaxed, positive energy where someone feels in sync with their surroundings or others. It’s often used to describe a harmonious, enjoyable atmosphere or connection.

Yesha said:

“OK.”

ChatGPT said:

Great! Anything else you’d like to add

You said:

“Give me a section about being in the zone.”

ChatGPT said:

Sure! “Being in the zone” refers to a state of intense focus and productivity where everything feels effortless. It’s often associated with peak performance, where time seems to fly by, and you’re fully immersed in the task.

You said:

“What is the academic source for that?”

ChatGPT said:

A classic source on the “flow” state, which is akin to being “in the zone,” is Mihaly Csikszentmihalyi’s book, Flow: The Psychology of Optimal Experience. This work explores the conditions and benefits of engaging deeply in an activity.

You said:

“Excellent. So give me two or three sentences on the nature of work in the future, which involves flow that stems from using the tools.”

ChatGPT said:
Certainly! In the future, work could be characterized by frequent “flow” states, thanks to AI tools that automate mundane tasks. This would allow people to focus on creative, high-level problem-solving, fostering innovation. Ultimately, this shift could lead to more fulfilling and productive work experiences.

Key Examples: Vibing Based on Data Types

  • Vibing with Text — tools like OpenAI Canvas (my favorite), Notion AI, and Claude Artifacts turn writing into a true back-and-forth process with AI. These tools help draft, revise, and reframe ideas instantly, making the act of writing faster, smarter, and more fun.
  • Vibing with images — tools like Midjourney, new OpenAi image tools, and Adobe Firefly enable rapid creation and iteration of visuals from text prompts. Designers now explore styles, concepts, and variations at unprecedented speed.
  • Vibing with video — AI tools like Runway, Pika, and Synthesia enable rapid video creation from text, voice, or static images. Storyboarding, editing, and even generating actors or scenes can now be done collaboratively with AI.
  • Vibing with music — tools like Suno, Aiva, and Udio let creators generate original compositions from text prompts or mood settings. AI collaboration allows Musicians to explore melody, genre, and structure in real time.
  • Vibing with voice — tools like ElevenLabs, Descript, and OpenAI.fm make it easy to generate, clone, or edit voices with stunning realism. From podcasts to narration, creators can now fine-tune tone, pacing, and emotion on the fly. (I’m a mega user of OpenAI voice pal — and I use my walks to explore new ideas with it).

The Downsides of Vibing

As usual, with great power come significant problems:

  1. Overconfidence – Feeling too “in the zone” can lead to overlooking mistakes.
  2. Lack of Critical Thinking – Flow can suppress questioning and reflection.
  3. Tunnel Vision – You might miss broader context or better alternatives.
  4. Burnout Risk – Constantly chasing that feeling may lead to exhaustion.
  5. Overreliance on Emotion – Decisions may become feeling-based, not fact-based.
  6. Exclusion of Others – Group vibing can unintentionally leave others out.

This initial set of problems is just the start — it reminds me of the (mis)joy of agile, which ignored longer-term considerations.

For a great review, see Claude Code saved us 97% of the work on the first try. Then, it failed utterly.

Actions for Leaders

On the organizational level:

  • Share the term “vibing,” its meaning, and its modalities.
  • Allow for encouraged individual vibing time (with open as well as focused goals)
  • Allow for small team vibing time where 2–3 people + AI work together.
  • Remember: Vibing does not replace architecture, rigor, testing, and robustness. In fact, vibing requires spending a lot of time on these non-ephemeral qualities.

On a personal level, allocate time to learning AI using the “T” method:

2. Video: Vibe Coding Is the Future (Source: Y Combinator, YouTube, 31.33m)

Source: Y Combinator, YouTube

“Andrej Karpathy recently coined the term “vibe coding” to describe how LLMs are getting so good that devs can simply “give in to the vibes, embrace exponentials, and forget that the code even exists.” We dive into this new way of programming and what it means for builders in the age of AI.”

Chapters (Powered by https://bit.ly/chapterme-yc) :

0:00 Intro

0:42 What is vibe coding?

1:00 What founders in the current YC batch are saying

4:35 Debugging and building systems

6:59 The models people are using now

10:01 What percentage of code is being written by LLM’s?

11:58 What changed, and what stayed the same?

18:08 How did Triplebyte candidate assessments, and how would that change in this era

21:37 Key skills that will remain relevant

23:01 How do you develop taste without classical training?

30:59 Outro

See the full video

3. New Model: Introducing Google Gemini 2.5 (Source: Yossi Matias, LinkedIn, 5m Read)

Source: LinkedIn

Key points from Yossi’s post (He is Vice President, Google. Head of Google Research.)

  1. Uses state-of-the-art reasoning — Gemini 2.5 are thinking models capable of reasoning through their thoughts before responding, resulting in enhanced performance and improved accuracy.
  2. Tops the LMArena leaderboard — which measures human preferences — by a significant margin
  3. Dominates on math & science benchmarks — Leads on GPQA, AIME 2025, and Humanity’s Last Exam (18.8% score without tool use), outperforming GPT-4.5 and Claude 3.5 Sonnet.
  4. With next-level coding intelligence, it excels at — creating visually compelling web apps and agentic code applications; on SWE-Bench Verified, the industry standard for agentic code evals, Gemini 2.5 Pro scores 63.8% with a custom agent setup
  5. Has long context window — Ships with 1 million tokens today (2 million coming soon), enabling comprehension of massive datasets, full codebases, and multimedia inputs.

Gemini 2.5 Pro is now available in Google AI Studio, and the Gemini app for Gemini Advanced users is coming soon with Vertex AI support.

Check out the full  review: Gemini 2.5: Our most intelligent AI model

See its comparison to other models:

Source: Google

4. New Image Tool: Open Introducing 4O Image Generation (Source: OpenAI, YouTube, 16m)

Source: OpenAI

“4o image generation rolls out starting today to Plus, Pro, Team, and Free users as the default image generator in ChatGPT, with access coming soon to Enterprise and Edu. It’s also available to use in Sora. For those who hold a special place in their hearts for DALL·E, it can still be accessed through a dedicated DALL·E GPT.

Developers will soon be able to generate images with GPT‑4o via the API, with access rolling out in the next few weeks.

Creating and customizing images is as simple as chatting using GPT‑4o – just describe what you need, including any specifics like aspect ratio, exact colors using hex codes, or a transparent background. Because this model creates more detailed pictures, images take longer to render, often up to one minute.“

Read all about it: Introducing 4o Image Generation | OpenAI

Formal openai video review w intro by Sam Altman: 4o Image Generation in ChatGPT and Sora

One cool review video: OpenAI Just Perfected AI Image Generation (See the Comparison!)

Conor Grennan says

I cannot overstate how revolutionary OpenAI’s new image generator is. But first, let’s take a moment to appreciate the weird, wonderful, modern-art-nightmares of the old version. Can I get an amen?” 

And adds his personal experiment:

Source: LinkedIn

5. MindLi Update: Feed Your AI FOMO With Our LinkedIn MindLi AI Group (Source: LinkedIn)

Source:LinkedIn

Explore our updated AI resources on our LinkedIn MindLi AI private group (Join Us).

Whether you’re revisiting familiar content or discovering something new, this stream is designed to keep you informed and ahead in the world of AI. Here’s a glimpse of the latest as of mid-March 2025:

  1. Jim Fan, NVIDIA Sr. Research Manager, talks about GR00T N1: The first open model for humanoid robots—transforming photons into actions with 30% improved performance. Open brain, open hardware. One token at a time.
  2. Etan Mollick, Associate Professor at The Wharton School, talks about “delve” and “xxx” as a giveaway for GPTs. 
  3. Kevin Misener, Senior Attorney – AI, Cybersecurity and Federal Contracts, talks about
    NIST releases AI 100-2 E2025: A definitive guide to AI/ML cybersecurity attacks and mitigations, paving the way for future standards
  4. Kai-Fu Lee, AI China Maven & CEO of 01.AI, talks about Deepseek’s most significant revelation: Open Source has won — making it hard for OpenAI to justify its price for just 1% better performance.
  5. Stephan Klein, Co-Founder & CEO of Curiouser.AI, talks about OpenAI shifts to Hype-as-a-Service™—because scaring the industry is more lucrative than developing AI.
  6. Amir Hershkovich, Founder @ Future AI | Creator of ChatBot Future.AI – AI Chatbots for Businesses, talks about how The AI revolution is already here—don’t get left behind. Learn, adapt, and embrace the future before it passes you by.
  7. Hanan Zakai, VP of Customers & GenAI Lead, discusses Considering deploying GenAI. Avoid starting with a data lake—prioritize solving real problems first.
  8. Linas Beliūnas, Financial Technology | AI, talks about China’s Manus AI has arrived—an unstoppable, viral AI agent ready to surpass DeepSeek and OpenAI.
  9. Elena Gurevich, AI Policy-Curious Attorney | AI Legal Strategy, Governance & Compliance, talks about GPT-4.5 here—OpenAI’s most powerful and emotionally intelligent model to date. However, it’s not recognized as a frontier model.
  10. Reuven Cohen, Agentic Engineer / aiCTO / Consultant, talks about Meet Mercury Coder: the fastest coding model globally, reaching 1,000 tokens per second with innovative language diffusion technology.

 

How to Keep in Touch

Wish to Stay Updated? 

For updates on DigitalRosh activities and news, sent by the group manager only. 
Join our WhatsApp group

Share Your Thoughts With Us

We will be happy to hear from you with any feedback, ideas, or questions.

Contact us.

About Leadership CONNECT

MindLi CONNECT, is a weekly source of news and inspiration for thinking aimed at DigitalRosh members and MindLi users. 

In every issue, you will find curated content from different industries. It would take you about 10 minutes reading to get a brief overview of the presented topics.

Want to delve deeper? You would need to have a subscription to the DigitalRosh site which is an affiliate site of MindLi. It is Free. Click on the links in each item or follow the breadcrumbs to the DigitalRosh website.

Enjoy!

See past CONNECT editions