This week was pure tech chaos: Google unveiled its smartest AI ever, the Internet suddenly collapsed, while the world was having an “emotional breakup” with punctuation. A lot happened at once — some thrilling, some annoying, all worth unpacking.

In Today’s AI Simplified: 

• Founder’s Insight: Early-stage sales works best when you treat it like science, not art. What is my take on this?
• AI News: Internet meltdown, Google launches an advanced model, and Big Tech warns the AI boom may echo the 90s bubble.
• AI Spotlights: A LinkedIn-growth tool, a punctuation drama that’s been in the spotlight for a long time seems to be resolved — the em dash, and a webinar that might change how you build automations.

Is sales in a startup an art, a science, or a craft?

In the early days, I think it’s mostly science.

Science means:

• You set a hypothesis: ICP, segment, problem, pitch, channel.
• You run experiments: different messages, offers, pricing, formats, geos, personas.
• You measure everything: reply rates, call-to-demo, demo-to-trial, trial-to-paid, sales cycle length.
• You question the “axioms” of your industry instead of copying them:

Do we really need SDR → AE → CSM?

• Does our ICP look the same as everyone else’s?
• Are we sure this must be sold via demos and not self-serve?

At this stage, sales is less about “charisma” and more about building a system that learns.

You’re not closing deals, you’re validating or killing hypotheses.

Craft come later – when you already know who to sell to and why they buy, but i never think that sales is an art.

But until then, if you treat sales as science and run it like a lab,

you’ll find truth much faster than if you treat it as magic.

1. Google Launches Gemini 3, Its Most Advanced Model Yet

Google just introduced Gemini 3, calling it its most intelligent and capable model so far. The new release marks a major step forward in reasoning, multimodal understanding, and autonomous agent behavior. It's rolling out across Search, the Gemini app, AI Studio, Vertex AI, and a new developer platform called Google Antigravity.

Gemini 3 Pro, available today in preview, shows significant improvements across virtually every benchmark. It handles deeper context, gives more direct and insight-rich answers, and can work across text, images, audio, video, and code with a 1M-token context window. 

Google is also introducing Gemini 3 Deep Think, an enhanced reasoning mode that outperforms even the base model on the most difficult evaluation tests and will come to Ultra subscribers after additional safety review. More models in the Gemini 3 series are coming soon.

2. Major Cloudflare Outage Disrupts X and Most of the Web

A widespread Cloudflare outage on Tuesday temporarily knocked major websites offline, including X, OpenAI, Letterboxd, and parts of League of Legends. Users saw “internal server error” messages tied to Cloudflare’s network, and even outage tracker Down Detector struggled to load due to the same underlying issue.

The disruption began around 11:30 a.m. UK time, with Cloudflare acknowledging the issue minutes later. Early updates indicated widespread 500 errors impacting multiple customers, the Cloudflare dashboard, and the company’s API. While Cloudflare didn’t immediately identify the cause, later statements said services were gradually recovering, though higher-than-normal error rates might persist during remediation.

Cloudflare, which sits behind a massive portion of global internet traffic, typically operates invisibly — handling security, routing, and uptime for millions of sites. Outages of this scale echo the AWS incident a month earlier, where a single provider issue cascaded across unrelated services. By early afternoon, some major platforms had begun loading again, though Cloudflare had not yet declared the incident fully resolved.

So if your website (or a site you tried to visit) blinked out today, chances are it wasn’t you. It was Cloudflare taking an unplanned break.

3. Google’s CEO Says the AI Boom Has “Irrationality” Built In

In an exclusive BBC interview, Alphabet (Google’s parent company) CEO Sundar Pichai compared today’s AI momentum to the 1990s internet boom — rapid innovation paired with heavy, uneven investment. He also outlined Alphabet’s expansion plans in the UK, the growing energy demands tied to AI development, and how major professions will adapt as AI tools become standard.

Key highlights

• The current AI boom includes “irrationality,” and no company is immune to a potential market correction.
• Alphabet’s market value has doubled in seven months to $3.5tn, driven by its full-stack AI approach (chips, models, data centers, and cloud).
• The company is investing £5 billion in UK AI infrastructure over the next two years and will begin training models locally at its DeepMind unit in London.
• AI development is driving up energy demand; AI already consumed about 1.5% of global electricity last year, according to the International Energy Agency.
• Pichai expects job transitions rather than widespread elimination, with success going to those who learn to use AI tools effectively. Professions like teaching and medicine will evolve, not disappear.

Tool of the week

Taplio — Social media assistant for LinkedIn

Taplio helps anyone looking to grow their presence on LinkedIn. It combines AI-generated content, scheduling, and analytics into one workspace, helping founders, creators, and teams stay consistent, understand what’s working, and turn LinkedIn into a real acquisition channel rather than a chore.

Key Features:

• AI LinkedIn writer: Generates post ideas and drafts tailored to your audience.
• Smart scheduling: Automates publishing at optimal times for reach and engagement.
• Analytics dashboard: Tracks post performance, follower growth, and engagement patterns.
• Competitor insights: Surfaces how peers and influencers are performing to guide your strategy.
• Profile optimization: Offers actionable suggestions to strengthen your LinkedIn presence.

Social Buzz

The Em Dash Era Is Finally Over?

Sam Altman’s post announcing that ChatGPT will finally stop sneaking in em dashes (if you tell it to in Custom Instructions) set off a wave of memes, grammar debates, and light existential panic across Reddit and X.

Across platforms, users test the “fix”, arguing over why LLMs latched onto the em dash in the first place. Top comments floated theories that it was a training artifact, an accidental “watermark,” or even a “poison pill” to stop AI text from blending too easily with human writing. Others joked that the internet just lost its favorite “AI tell,” and a few insisted the fix still doesn’t work. And then there were the em-dash loyalists, defending the punctuation as a perfectly normal writing choice long before ChatGPT ever existed.

Altman’s announcement on X went viral instantly. Most replies treated it like an oddly poetic milestone, OpenAI solving the internet’s pettiest annoyance. Overall: half relief, half comedy, and a persistent worry that AI text just got a little harder to recognize.

AI Event Pick

Building With Claude in Europe

Hosted by Anthropic & MeshAI

📅 Date: November 25, 2025

🕙 Time: 10:00 am GMT

Anthropic is hosting a live webinar focused on helping teams understand and build AI agents with Claude. The session walks through what “agency” means in practical terms, how agentic systems differ from traditional workflows, and how to design, test, and deploy agents safely. Attendees will see live demos, hear real deployment examples from European organizations, and learn the fundamentals needed to bring Claude-powered agents into production.

Register here.