Welcome to your weekly AI Newsletter from AITechCircle!

I’m building and implementing AI solutions and sharing everything I learn along the way…

Check out the updates from this week! Please take a moment to share them with a friend or colleague who might benefit from these valuable insights!

Today at a Glance:

• Actions for Organizational Leaders to Set a Steady AI Race
• Generative AI Use cases repository
• AI Weekly news and updates covering newly released LLMs
• Courses and events to attend

Is the AI race Your organization’s AI Race?

My work with customers across the region and conversations with organizational leaders kept reminding me that two AI races are going on in the AI industry:

​

1. The first is the AI vendor race, a whirlwind of innovation familiar to anyone in the field. New large language models (LLMs) and AI technologies emerge almost daily, making staying current challenging. This relentless pace was a key motivator for writing this weekly newsletter: to gain actionable insights and track noteworthy trends for myself and the wider community.
2. The second race is the organizational AI race – the strategic effort to determine how, where, and why to adopt AI within a business. The vendor race profoundly influences this race, as the rapid evolution of tools shapes the possibilities and pressures leaders face.

This was very well articulated in the Gartner’s keynote, ‘Pacing Yourself in the AI Races,‘ at the Gartner IT Symposium/Xpo that was also covered in this weekly newsletter during early 2025 ‘Race to Deliver Generative AI Outcomes‘

The core message?

Organizations don’t need to sprint alongside the vendors. Instead, they can set a steady pace, aligning AI adoption with their unique goals, resources, and challenges. ​​

But how can leaders ensure their AI race is deliberate and effective?

Here are actionable steps to establish a steady, organization-specific AI race:

Inspite of last 2 years of AI race, large number of organizations are just going through very silos approachs of taking MVPs, POCs of early Use cases to adopt Generative AI.

Today, let’s take just two actions to take in your organization:

Define Your “Why” for AI Adoption

• Action: Start by clarifying the business outcomes you aim to achieve, whether it’s boosting employee productivity, optimizing processes, or reimagining your business model. Avoid adopting AI just because it’s the latest trend. In other words, it’s called defining an AI Strategy. You can refer to our earlier resources, ‘Building a Game-Changing AI Strategy: Step-by-Step Guide and Exercises for Your Organization’ to get started.
• Why It Matters: A clear purpose keeps your efforts focused and prevents wasted resources on mismatched tools or overhyped solutions.

Assess Your Current Capabilities

• Action: Conduct an internal Generative AI maturity Assessment of your data, AI infrastructure, workforce skills, and technology stack. Identify gaps. Here is the resource ‘Generative AI Maturity Framework for Structured Guidance‘ to do the maturity assessment for your organization.
• Why It Matters: A steady race requires a solid foundation. Gartner emphasizes that AI success depends on “AI-ready” data and security, so address these fundamentals before accelerating.

Call for Action:

The AI vendor race will keep churning, but your organization’s race doesn’t have to mirror it. By setting a deliberate pace, the AI race truly is your organization’s AI race.

​

Weekly News & Updates…

Last week’s AI breakthroughs marked another leap forward in the tech revolution.

1. X has launched Grok-3 beta, and it is trained on Colossus supercluster with 10x the compute of previous state-of-the-art models. Grok 3 displays significant improvements in reasoning, mathematics, coding, world knowledge, and instruction-following tasks
2. Mira Murati, former CTO of OpenAI, has started Thinking Machines Lab with several other prominent engineers.
3. BioEmu-1 from Microsoft Research. This deep learning model can generate thousands of protein structures per hour, unlocking new possibilities for protein scientists and drug discovery and research. link​
4. Microsoft introduced Majorana 1, the world’s first quantum chip powered by a new Topological Core architecture that it expects will realize quantum computers capable of solving meaningful, industrial-scale problems in years. It uses the world’s first topoconductor, a breakthrough type of material that can observe and control Majorana particles to produce more reliable and scalable qubits, which are the building blocks for quantum computers
5. Clone Robotics has announced Protoclone, the world’s first bipedal, musculoskeletal android. The Protoclone is a faceless, anatomically accurate synthetic human with over 200 degrees of freedom, 1,000 Myofibers, and 500 sensors. link​
6. Figure Robot has launched Helix, a generalist Vision-Language-Action (VLA) model that unifies perception, language understanding, and learned control to overcome multiple longstanding challenges in robotics. link​
7. NEO Gamma is the next generation of home humanoids designed and engineered by 1X Technologies. link​

The Cloud: the backbone of the AI revolution

• Cloud AI Resources Where You Need Them: Announcing NVIDIA L40S GPU on Oracle Compute Cloud@Customer and Oracle Private Cloud Appliance link​
• Telenor Builds Norway’s First AI Factory, Offering Sustainable and Sovereign Data Processing link​

Generative AI Use Case of the Week:

Several Generative AI use cases are documented, and you can access the library of generative AI Use cases. Link​

Favorite Tip Of The Week:

Here’s my favorite resource for the week.

​TogetherAI has published the DeepSeek-R1 Quickstart, which covers:

• Applications and use cases
• A complete guide to prompting R1
• Reasoning vs non-reasoning models
• Model limitations

Potential of AI:

OpenAI has launched SWE-Lancer, a new, more realistic benchmark for evaluating the coding performance of AI models. The benchmark includes over 1,400 freelance software engineering tasks from Upwork.

SWE-Lancer tasks span the full engineering stack, from UI/UX to systems design, and include a range of task types, from $50 bug fixes to $32,000 feature implementations. SWE-Lancer includes independent engineering and management tasks, where models choose between technical implementation proposals.

Things to Know…

DeepSeek is running OpenSource Week, and until now, you can have a six-day release that drives AI.

• ​Day 1: FlashMLA – an efficient MLA decoding kernel for Hopper GPUs.
• ​Day 2: DeepEP – the first open-source EP communication library for MoE model training and inference.
• ​Day 3: DeepGEMM – an FP8 GEMM library that supports dense and MoE GEMMs, powering V3/R1 training and inference.
• ​Day 4: DualPipe – a bidirectional pipeline parallelism algorithm for computation-communication overlap in V3/R1 training. Expert Parallelism Load Balancer (EPLB) – Different experts are assigned to different GPUs when using expert parallelism (EP). Because the load of different experts may vary depending on the current workload, it is essential to keep the load of different GPUs balanced. Profiling Data in DeepSeek Infra – Analyze computation-communication overlap in V3/R1
• ​Day 5: Fire-Flyer File System (3FS) – a parallel file system that utilizes the full bandwidth of modern SSDs and RDMA networks
• ​Day 6: DeepSeek-V3/R1 Inference System Overview

The Opportunity…

Podcast:

• This week’s Open Tech Talks episode 154 is “Generative AI Risks and Governance: What Business Leaders Need to Know with Terry Ziemniak.” he is a Fractional CISO and Partner at TechCXO

Apple | Spotify | Amazon Music ​

Courses to attend:

• ​Introduction to LangGraph from Langchain Academy

Events:

• ​Dubai AI Festival 2025, April 23-24, 2025, Dubai, UAE

Tech and Tools…

• ​olmOCR – A toolkit for training language models to work with PDF documents
• ​VisionAgent is a library that helps you utilize agent frameworks to generate code to solve your vision task

And that’s a wrap!

Thank you, as always, for taking the time to read.

I would love to hear your thoughts. Please reply and share what you found most valuable this week. Your feedback means a lot to me.

Until next week,

Kashif Manzoor