Servers Servers Servers

Servers

Google’s servers, 1999

In 1999, Google's servers were ten CPUs across four servers. 

Google doesn’t publicly disclose the exact number of servers it operates, but estimates suggest it’s in the millions.  A 2011 analysis based on Google’s energy usage suggested around 900,000 servers, but that was over a decade ago. In 2016, Gartner estimated Google had around 2.5 million servers globally. Google has designed systems like Spanner to manage fleets ranging from 1 million to 10 million machines, hinting at massive infrastructure growth. Given the scale of Google’s services—Search, YouTube, Gmail, Cloud, and AI—it’s reasonable to assume the number is well above 2.5 million today, spread across dozens of data centers worldwide.

Other tech giants like Amazon and Microsoft don't publish exact server counts, so these are educated estimations. 

Amazon (AWS) probably has ~4–5 million  as the largest cloud provider globally with a massive scale across 100+ data centers.

Google probably has ~2.5–3 million+ for its heavy investment in AI and search, and is expanding rapidly with Gemini and Cloud.

Microsoft (Azure) is close with ~2–3 million+ but has a strong enterprise presence and is growing with Office 365, Copilot, and Azure AI.

 

AI-Powered Browsers

AI-powered browsers are reshaping how we interact with the web. They are moving us from a more passive browsing to hopefully more intelligent browsing by using these task-completing companions.

The 3 big names are:
Microsoft Edge, which uses ChatGPT (OpenAI) in its Copilot. The sidebar assistant is for writing, summarizing, and answering questions across sites
Google Chrome has its own Gemini AI for enhanced search, writing help, and experimental tab organization.
Opera One is not as popular, but its Aria (also powered by OpenAI) for in-browser assistance for coding, content generation, and web Q&A.

There are also some smaller, less well-known browsers using AI, such as Brave, Arc Browser, Perplexity Comet etc.

The underlying technologies vary. Large Language Models (LLMs) are used by most browsers using models like GPT-4, Claude, or Gemini to power natural language understanding and generation.

Agentic AI browsers, like Comet and Dia, go beyond chat and perform multi-step tasks like booking flights or managing emails autonomously.

Advanced browsers maintain memory across tabs and sessions. This context awareness enables smarter comparisons and task continuity.

Brave’s Leo and SigmaOS emphasize local processing and anonymity, avoiding cloud-based data sharing, and these privacy enhancements should be encouraged across all browsers. As these browsers become more autonomous, they face risks like prompt injection attacks, where hidden commands in web content can trick the AI into executing unintended actions.

Podcasting and iTunes U

iTunes U user at NJIT

Podcasts at Apple are turning 20 years old, but Apple officially discontinued iTunes U at the end of 2021. The platform, which launched in 2007, was designed to help educators distribute lectures, assignments, and other learning materials through the iTunes ecosystem. It was especially popular with universities and K–12 institutions - for a time.

As Apple shifted its focus to newer tools like Classroom and Schoolwork, which offer more interactive and integrated experiences for iPads and Macs, iTunes U gradually became outdated. Apple announced the phase-out in 2020, giving educators time to transition to an alternative platform. That was partially why there was a temporary bump up in schools participating in MOOCs. 

NJIT iTunes U

I was at NJIT 20 years ago and spearheaded the university's inclusion in the initial "sweet 16" schools in iTunes U. It was important in its time, and I would update the new schools as they were added. I was disappointed in Apple's lack of support for the academic version of podcasting and was further disappointed when they dropped it after 14 years, but it had been losing its lustre. iTunesU (and iTunes itself) didn't really move forward, and the newer apps for educators allowed us to guide students through lessons in real time, share assignments, and track progress more efficiently. 

I searched for some of those old Serendipity35 iTunes U posts. Very nostalgic in 2025.

These are some of the general online learning platforms where some of that content was repurposed.
Khan Academy – Free, nonprofit platform offering lessons in math, science, economics, and more. Great for K–12 and beyond.
Coursera – Offers courses from top universities like Stanford and Yale. Many are free to audit, with paid certificates available2.
Udemy – A massive marketplace for courses on everything from coding to photography. Often discounted and self-paced.
FutureLearn – UK-based platform with university-led courses and microcredentials.
Institutional Tools
Canvas LMS and Blackboard Learn – Widely used by schools and universities for managing coursework, assignments, and communication.
Moodle – Open-source learning management system used by many institutions worldwide.
Open Educational Resources
Wikiversity and Wikibooks – Free educational content created and curated by volunteers. Great for self-learners.
Classroom & Student Engagement
ClassDojo – Focuses on classroom behavior and communication, especially for younger students.
Remind - Helps teachers communicate with students and parents via messaging.
Lumio – Turns lessons into interactive, collaborative experiences.

 

Going Viral

Reading Amber Mac's newsletter this week, I saw a connection with something I was writing on one of my other blogs about this idea of having content "go viral." It sounds like a great thing. But does it come at a cost?

Yes, it's tempting to jump on TikTok trends and participate in the bite-size-ification of social, but it's also possible to produce a steady and consistent flow of valuable, high-quality content that never goes viral. Don't get me wrong, I'm happy to see our video views rise, but it's not essential that we trend. I think we all know how that game works. Say something controversial or do something outrageous, and the algorithms will thank you for it.

In my opinion, that spike in attention might be a win, but it's a short-term win. In other words, not everyone should go viral. If that's what you're chasing, it could be a soul-crushing experience where you lose yourself and your authentic community along the way.  - Amber Mac

viral

Bloggers often want their posts to go "viral." The word “viral” made a leap from medical terminology to broader cultural usage in the late 20th century, particularly in the context of marketing and media. I assumed that it was social media that moved the term from medical usage, but it is actually a bit earlier than the explosion of social media. In 1989, The Oxford English Dictionary cites the earliest use of “viral” to describe the rapid spread of information, marking its first known non-medical usage.

The term gained traction in marketing in 1999, especially “viral marketing,” which described campaigns that spread quickly and organically—much like a virus. In the early 2000s, phrases like “going viral” and “viral video” emerged. and by 2004, “going viral” was used to describe content that rapidly gained popularity online. From 2009 onward, viral became mainstream, fueled by the rise of social media platforms and shareable content.

It is a good and powerful metaphor. Like a biological virus, digital content can replicate and spread uncontrollably. That semantic link made “viral” the perfect word to describe the phenomenon of explosive online popularity.

I don't think I have had a "viral post," though I have had some posts that seem to get more views over the years than most of mine. But "viral" is when the surge of views hits all at once.

Can you push a post into the land of viral? I don't think so, but you can find articles about "how to," like 21 key elements for viral blog posts or explore 32 proven tactics to boost your chances - but there are certainly no guarantees.

Those kinds of articles will suggest things to do like these: Know your audience and tailor your tone, topic, and style to what resonates with them. Tap into trending issues, emotional stories, or highly useful how-tos. Use eye-catching images. Share across platforms with tailored captions and hashtags. Use analytics to see what’s getting attention and tweak accordingly. All of those make sense, and I tend to employ them most of the time.

Some suggestions probably do increase your chances of viality, but go against my own blogging philosophy: Make it skimmable. Short paragraphs, bullet points, and bolded key phrases. Engage influencers. Reach out to people with large followings who might share your post. I don't do email marketing, other than if someone decides to follow my blog via email notifications.


"Viral" is hardly the only medical term that is now in broader usage. If you're curious about that, check out this other blog.