Work Less: AI at Work

This week on Work Less, we’re talking about $180 million in fresh funding for n8n and what it means for the automation wars - can they really take on Zapier, Make, and the old-school RPA giants?

Then we test drive OpenAI’s brand new browser “Atlas by ChatGPT” (spoiler: it’s… not great), and unpack what makes AI browsers like Daya and Perplexity different, or at least less terrifying for your passwords.

Finally, we get philosophical (and a little doomsday-y) about “AI slop” — the growing mess of machine-generated junk clogging the internet, and whether the “dead internet theory” might be less of a conspiracy and more of a preview.

No hype, no fluff - just two people trying to make sense of the automation chaos.

🧰 Tools, Companies & Topics Mentioned

• n8n (Series C funding, $180M raise)
• Sequoia Capital, Accel, Nvidia (investors)
• Zapier, Make (Integromat), Lindy, Relay, Gumloop
• OpenAI’s Atlas Browser
• Daya, Arc (The Browser Company), Perplexity
• Claude 4.5 Sonnet (Anthropic)
• Gemini (Google)
• ChatGPT (GPT-5)
• Claude Skills
• AI Slop / Dead Internet Theory
• Kurzgesagt (video reference)

This week, Justin and Tom unpack the chaos around OpenAI’s new Agent Builder, the so-called “Zapier killer.” Spoiler: it’s not killing anything yet (except maybe developer patience). We break down what Agent Builder actually does, how it leans on the MCP protocol, and why it’s nowhere near replacing tools like Zapier, Make, or Lindi for real-world, multi-user automations.

Then, we shift gears to talk about Retool’s quiet domination of the agent-building space, their slick new AI + workflow features, and why their version of “vibe coding” might finally make sense. Airtable’s Omni Builder and Replit get some side-eye too, all in a week where OpenAI tries to hug every industry at once.

Also: guardrails, evals, why constraints make for better creativity, and a quick nod to Canadian Thanksgiving 🇨🇦.

🧰 Tools & Companies Mentioned:

• OpenAI Agent Builder
• Zapier
• Make / Integromat
• Lindi
• Retool
• Airtable Omni Builder
• Replit
• Lovable
• MCP (Model Context Protocol)
• Gemini (Google)

This week we get into the weeds of AI hype vs. reality:

• Justin brings up Higgs Field, an AI-first marketing platform with the most Gen Z-coded UX imaginable. Messy, loud, and -surprisingly- brilliant for marketers.
• Tom unpacks how Oracle added a casual $1 trillion to its market cap by shifting from ERP giant to AI infrastructure heavyweight (plus why some companies are still running the other way).
• We break down the bizarre three-sentence “joint statement” from Microsoft + OpenAI that somehow manages to say… absolutely nothing.

Also: AI influencers, fake UGC everywhere, and whether we’ll ever trust what we see online again.

If you want the no-fluff version of what’s actually happening in AI, this one’s for you.

Companies & Tools Mentioned

• Higgs Field (AI marketing platform)
• Anthropic, OpenAI, Google (model providers behind Higgs Field)
• V0 (AI video)
• Nano Banana (AI images)
• Oracle (ERP turned AI infra player)
• SAP (ERP competitor)
• AWS, Google Cloud, Azure (cloud providers in the mix)
• Stargate (Oracle + OpenAI + SoftBank mega project)
• Microsoft (partner drama with OpenAI)
• OpenAI + Jony Ive + LoveFrom (hardware rumors)
• Skydance & Paramount (Larry Ellison family business moves)

This week we dig into a new survey on AI and workers from Kyla Scanlon that asks the big question: do employees actually trust their employers to use AI in their favor? Spoiler: real estate agents and Hollywood don’t. Healthcare somehow does.

We also unpack the messy state of AI training at work (hint: most people aren’t getting any) and why “upskilling” is the thing employees want most, even if it means learning the tools that could replace them.

Then we shift gears into tech drama: Atlassian just dropped $610M in cash on The Browser Company (makers of Arc). Why would a project management giant buy a browser? Two words: orchestration layer. Everyone from Zapier to OpenAI is fighting to be the conductor of the SaaS orchestra, and browsers might just be the new front line.

Oh, and we end up talking about Alien (the new TV show) and Dexter because of course we do.

Companies & Tools Mentioned:

• Kyla Scanlon’s “AI That Works for Workers” survey → link
• Atlassian
• The Browser Company (Arc)
• Loom
• Help (ticketing system)
• Zapier
• Perplexity (Comet browser beta)
• OpenAI (rumored browser project)
• AWS
• Area, Trace (AI orchestration tools)

Is voice the real future of AI? Tom and Justin dig into Assembly AI, Whisper Flow, and why typing might already be outdated. We also get into the banana-shaped elephant in the room: Google’s new Nano Banana model, a tiny but scary-good image editor that could replace half of Photoshop.

But here’s the kicker: MIT says 95% of AI projects fail. We break down why, which departments actually should automate, and which ones you should absolutely keep away from AI (spoiler: don’t trust it with your finance team).

If you’ve ever wondered whether AI should talk to your customers, design your ads, or just stay in its lane, this one’s for you.

🛠️ Tools & Companies Mentioned

• Assembly AI - voice-to-text + AI phone agents
• Whisper Flow / Super Whisper - hands-free dictation apps
• Eleven Labs - voice generation
• Nano Banana (Google) - new image model with scary good consistency
• MIT - study on AI project failure rates

Justin and Tom dig into the growing exec–team AI gap, and what actually works to close it. 

From exec hackathons (yes, that's a thing now) and weekly “show your prompts” rituals, to putting bumpers on shadow IT before someone wires a weekend AI toy to your codebase, it’s a playbook for making AI normal at work. 

We also side‑eye Apple’s latest Siri/Gemini rumor, explain why Notion Mail finally pried Justin off Superhuman, and kick the tires on Airbook, aka “cursor for analytics.”

What’s inside:

• Leaders who model the behavior win. Exec retreats turned into mini hackathons (with Replit) and “show your prompts” at all‑hands beat vague “use AI” mandates. Make it impossible not to use AI, set clear behaviors, cut red tape, and carve out time to tinker. 
• Kill shadow IT with a “happy path.” The AI era version isn’t someone using Asana instead of Jira, it’s giving an AI wrapper to your codebase and CRM. Define approved tools and data guardrails so folks can experiment in the open. Bumpers on, bowling starts.
• Apple’s Siri situation. Rumors of Gemini as an option signal Apple still shopping for a story, and maybe for a model. Interesting strategy, messy execution.
• App of the Week: Notion Mail. Justin finally jumps from Superhuman: similar shortcuts and clean UX, but better AI‑assisted labeling + flexible custom views. (Gmail only, for now.)
• Tool test: Airbook (cursor for analytics). Connects data, writes SQL from chat, spits out charts. Nice power‑user flow (you can even copy code to ChatGPT), but the credit model bites and it struggles with messy, inconsistent spreadsheets. Verdict: shines with structured data; do your data homework first.

Tools & companies mentioned:

• StockX (exec hackathon), Replit, Headspace, Lenny’s Newsletter/Podcast (25 adoption tactics) 
• Apple, Siri, Google Gemini
• Notion Mail, Notion Calendar, Superhuman, Gmail/Google Workspace
• Airbook (analytics), Fivetran, Google Analytics 4 (GA4), ChatGPT

This week, we unpack the drama around GPT-5, a release so incremental that some folks are mourning GPT-4 like it’s a lost pet. We dig into why OpenAI quietly pulled the old models, the emotional attachments people have to LLM quirks, and why the thinking mode might be the only real upgrade worth caring about.

We also get into:

• The hidden superpower of GPT-5’s massive context window (and why it matters more for devs than casual chatters)
• Pricing drops that could make Nano-powered AI features the next SaaS gold rush
• Why Anthropic’s vertical-specific models might dodge alignment headaches
• Whether we’ve hit an innovation plateau, and which lab might break through first
• The weird moment where Tom got genuinely mad at an AI for cropping images wrong

By the end, you’ll either be cautiously optimistic or ready to join the “bring back my model” movement.

In this week's episode, we break down:
Why Slack’s AI summaries are hilariously ineffective

•  Our hands-on review of Dia, the new AI-first browser
• Early impressions of Arc Browser's controversial pivot to an AI browser, Dia
• Google's strategy to compete with the new "Web Guide" AI search and SEO's uphill battle
• Thoughts on Meta’s AI (uninspiring) ambitions with personalized AI glasses

___________________________

Follow Tom:
• XRay: https://xray.tech/
• LinkedIn: linkedin.com/in/tomnassr   

Follow Justin:
• Switchboard: https://www.withswitchboard.com/
• LinkedIn: linkedin.com/in/wattjustin    
• Twitter: twitter.com/just_watt

___________________________

Tools & Companies Mentioned:
• Slack (AI summaries, Huddles)
• Arc Browser
• Dia Browser (Arc's AI-first replacement)
• Notion
• Airtable
• Salesforce
• Anthropic
• OpenAI

In this episode, we cut through the AI agent hype flooding LinkedIn feeds to reveal what's actually working for small and mid-market businesses today.

We talk about:

• Current AI agent capabilities versus marketing promises (00:00)
• How reasoning models have changed the agent landscape (03:13)
• Code vs. no-code approaches for different team structures (07:15)
• Model flexibility in agent workflows (10:36)
• Practical security considerations for implementation (17:39)
• Top agent-building tools worth your attention (22:25)
• Grok3's launch analysis and business implications (28:21)
• Microsoft's quantum computing breakthrough (32:44)

Plus, we break down context window limitations and how to build an effective agent strategy using the right models for specific tasks.

__________________________

Follow Tom:
• XRay: https://xray.tech/
• LinkedIn: linkedin.com/in/tomnassr   

Follow Justin:
• Switchboard: https://www.withswitchboard.com/
• LinkedIn: linkedin.com/in/wattjustin    
• Twitter: twitter.com/just_watt    

This week we're exploring how AI is reshaping business interactions through Boardy's innovative approach to networking. We also chat about OpenAI's Deep Research and Google Gemini 2.0's release.

Learn about:
• How Boardy raised $8M using its own AI networking tool
• Converting conversations into structured business data
• Internal vs. external uses of conversational AI
• Building tools that meet users where they are
• The latest capabilities of Google's Gemini 2.0
• Real-world cost comparisons of different AI models


Follow Tom:
• XRay: https://xray.tech/
• LinkedIn: linkedin.com/in/tomnassr    

Follow Justin:
• Switchboard: https://www.withswitchboard.com/
• LinkedIn: linkedin.com/in/wattjustin    
• Twitter: twitter.com/just_watt    

Tools mentioned:
• Boardy: https://www.boardy.ai/
• OpenAI Deep Research: https://openai.com/index/introducing-deep-research/
• Gemini 2.0: https://blog.google/technology/google-deepmind/gemini-model-updates-february-2025/


Chapter Markers:
00:00 - Intro: Boardy & Recent Viral Moment
05:20 - What is Boardy?: The AI Super Connector Explained
08:29 - Rethinking Business Tools with AI
12:41 - Structured Data & Architecture Behind Conversational AI
15:08 - Is Boardy an AI Agent? Breaking Down the Buzzwords
17:17 - Real-World AI Agent Applications
22:47 - AI Model News & Updates
26:23 - Evaluating AI Models
30:48 - Palette Cleansers