Hugging Face Expands Robotics 🤖, A Letter to Tech Leaders 📩, OpenAI Builds A-SWE 🧑‍💻

A Letter to Tech Leaders: Beyond the AGI Mirage

📉 The Layoff Wave Reshaping the Tech Landscape

2025 has been a turbulent year for tech professionals. AI is transforming the workplace faster than anyone anticipated — but not always for the better.
According to Layoffs.fyi, over 23,500 tech employees have already lost their jobs this year across 93 companies. The list of companies making cuts includes both tech giants and rising startups.

• Microsoft is preparing for another wave of layoffs in May 2025, following cuts to 2,000 employees earlier this year and 650 in the Xbox division in September 2024. Internal shifts include efforts to boost the engineer-to-manager ratio and remove low performers (MSN News, April 2025).

• Google is also planning a fresh round of layoffs, citing AI-led restructuring and performance-based evaluations.

• Automattic (WordPress) laid off 270 employees, or 16% of its workforce, while Canva eliminated 10–12 technical writer roles as part of a shift toward AI-generated content.

• TikTok cut 300 jobs in Dublin — 10% of its Irish team — and Ola Electric let go of over 1,000 employees, marking its second major downsizing in just five months.

• Other major layoffs include Siemens (5,600 jobs), HP (2,000 jobs), Blue Origin (1,000+ engineers), and Salesforce (1,000 jobs).

• Meta kicked off the year by laying off 3,600 employees (~5% of its workforce), targeting "low performers" while continuing to hire for AI positions.

• Amazon reduced its communications team by several dozen employees, aiming to "streamline, speed up work, and bring teams closer to customers" (MSN).

The justification for many of these layoffs? A belief that AI will replace human labor, particularly in knowledge work. But this reasoning is built on shaky ground.

📬 To the Tech Leaders: A Cautionary Note on the AGI Hype

Dear Builders of the Digital World,

There’s a growing narrative in boardrooms and conference keynotes: that Artificial General Intelligence (AGI) — machines capable of doing anything a human can — is just around the corner.

Tech figures like Sam Altman, Dario Amodei, and Elon Musk have predicted AGI in a matter of years. Media voices like Ezra Klein and Kevin Roose have joined the call to "prepare society" (Bloomberg Opinion – Gideon Lichfield).

But let’s be honest: nobody agrees on what AGI even means.

Some define it as:

• AI that passes most human tests.

• AI that wins Nobel Prizes.

• AI that reasons and feels like a human.

• AI that physically operates in the real world.

Yet even today’s most powerful models — like ChatGPT, Gemini, and Claude — are still narrow in scope. As Gideon Lichfield explains, they often:

• Succeed at structured tasks (e.g., coding, drafting documents)

• Fail at ambiguity, ethics, creativity, and emotion

• “Hallucinate” false facts

• Require clearly bounded problems to function

In truth, AGI is a moving target, not a milestone. It’s a marketing term, not a scientific breakthrough.

By leaning on AGI to justify layoffs, we risk:

• Losing deep contextual expertise

• Breaking team cohesion and morale

• Overestimating AI’s actual capabilities in dynamic environments

🔍 Human intelligence is not general; it’s evolved and situational.

We are shaped by gravity, language, emotions, history, and purpose. No LLM can replicate that — and may never be able to.

So instead of using AGI as an excuse for mass terminations, let’s:

• Rethink org structures alongside AI augmentation

• Invest in training, not just automation

• Recognize human adaptability as our true competitive edge

💡 To the Workforce: Embrace the Tools, Not the Hype

Dear Builders, Makers, Thinkers, and Doers,

While AGI might be distant, one form of AI is already reshaping the way we work: Agentic AI.

Agentic AIs are systems that take action — not just give answers. Think of tools that can:

• Create and send follow-up emails

• Schedule meetings based on transcripts

• Fill out web forms

• Automate multi-step tasks across platforms

Products like Zoom’s upcoming agents, AutoGPT, and Devin are examples of this shift.

💪 These are tools you can start using today.

Rather than fearing job loss, build your job around these tools:

• Use Agentic AI to offload repetitive tasks.

• Learn prompt design, process chaining, and basic automation.

• Focus on what humans do best: judgment, leadership, vision, collaboration.

The more you engage with AI, the more valuable you become — not just to employers, but in building your own future.

🧭 A Future That’s Human-Led, AI-Powered

Here’s the truth we all need to embrace:

AI is not here to replace us — unless we choose to replace ourselves.

• To leaders: stop trimming teams blindly under the AGI illusion. Build thoughtful, AI-augmented organizations that empower people.

• To workers: step into this new landscape with confidence. Learn the tools. Lead the change.

In the age of Agentic AI, your best move isn’t to resist the future — but to shape it.

💬 OpenAI Developments

Scammers Use OpenAI API to Flood 80,000 Websites With Spam

AkiraBot, an AI-powered spambot using OpenAI’s GPT-4o-mini, targeted over 80,000 small and medium-sized business websites with custom spam comments promoting fake SEO services. It used tactics like CAPTCHA evasion, proxies, and site-specific messages to avoid detection, prompting OpenAI to disable the associated API key and investigate further.

OpenAI Is Building A-SWE, An AI That Does Everything A Human Software Engineer Does

OpenAI is developing a fully autonomous AI coding agent called "Agentic Software Engineer" (A-SWE), designed to handle the entire software development lifecycle—from writing code and QA to documentation—far beyond current tools like GitHub Copilot. With OpenAI's top-tier models and massive user base, A-SWE could significantly disrupt the software engineering industry and potentially redefine developer roles by the end of 2025.

🚀 Tech Industry Moves

NVIDIA is partnering with major manufacturers to produce its AI supercomputers and Blackwell chips entirely in the U.S., with facilities in Arizona and Texas expected to scale up production over the next 12–15 months. This initiative aims to strengthen supply chain resilience, create hundreds of thousands of jobs, and generate up to half a trillion dollars in AI infrastructure using NVIDIA’s own AI, robotics, and digital twin technologies.

Meta is beginning to train its AI models using public posts and comments from adults in the EU, as well as user interactions with Meta AI, to better reflect European cultures, languages, and nuances. Users will be notified and given the option to object, and Meta emphasizes that its approach complies with EU laws and excludes private messages and data from minors.

Anthropic researchers are studying how AI models like Claude generate rhyming lines, handle hallucinations, and reason through tasks, revealing that hallucinations often occur when the model recognizes a name but lacks factual information, leading it to fabricate responses. Their AI microscope project is an early-stage interpretability tool aimed at uncovering how models make inferences, though it currently only works on short prompts and limited aspects of model behavior.

Google is testing a new AI Mode launch method on Android by placing a prominent "Ask AI Mode" circle directly in the Search field, replacing the usual voice and Lens icons. This update includes Google Lens integration for image input and simplifies the interface with shortcuts for voice search, Lens, and gallery access, while iOS users can now launch AI Mode via a homescreen widget.

Microsoft has begun rolling out its controversial Recall feature, which takes continuous screenshots to help users search past activity, emphasizing local storage, opt-in use, and user control over data. Despite privacy upgrades and reassurances, experts remain concerned about potential misuse, especially since earlier versions were shown to be vulnerable to data extraction by ethical hackers.

Hugging Face has acquired French humanoid robotics startup Pollen Robotics, known for its open-source robot Reachy 2, to expand its robotics efforts. The move follows prior collaboration between the two on “Le Robot” and reflects Hugging Face’s commitment to making robotics more accessible to developers.

💡 AI Tools

HiDream
HiDream is an AI-powered video generation tool that allows users to create high-quality short videos from simple text prompts or concepts. It’s ideal for creators looking to bring visual stories to life quickly with minimal technical skill.

UNO
UNO is an AI operating system developed by Rabbit that functions as a universal interface, enabling users to control apps and services through natural language. It acts like an AI-powered assistant that learns your habits and simplifies task execution across devices.

WordPress AI
WordPress AI integrates artificial intelligence directly into the WordPress platform, assisting users in content creation, SEO optimization, and image generation. It streamlines website management, making it easier for bloggers and businesses to maintain engaging, AI-enhanced websites.

Fantasy Talking
Fantasy Talking is an AI app that animates characters or photos into realistic talking avatars using voice input or text. It’s popular for creating personalized messages, storytelling, or entertainment content with lifelike facial expressions and lip-syncing.

👥 Connect & Feedback!

👉 Join Us:

📧 Advertise In Weekly AI News:

📧 Contact directly at [email protected]

😍 Share with your friends!