Tom Solid | AI Productivity

2.5 million people canceled ChatGPT last month. But most of them just swapped logos. Same prompts. Same "write me an email" usage. I built a 30-person AI team with Claude instead. The tool switch isn't the win. The system behind it is. Full breakdown in this video.

If you've written off AI because your first attempt didn't work, that's normal. Everyone's first attempt is bad. The question isn't whether you're technical enough. It's whether you've found the right person to learn from. And if you're here, you might have. #AI #Productivity

I've seen people with zero coding background set up working AI workflows in an afternoon. Not because they're geniuses. Because someone walked them through it step by step. The difference between "too hard" and "surprisingly doable" is almost always just a guide.

"The skill ceiling is too high." I got this comment from someone who watched my video and decided AI tools aren't for regular people. My reply: "The skill ceiling is lower than you think. It's just that nobody's shown you yet." Let me explain what I mean.

@emollick Vibework expanding beyond coding is huge. But coding has a clear test: does it run? Design has no binary check. People who get the most out of Stitch will be those who already know what they want. Without that clarity, AI just produces beautiful noise.

I think Google's new Stitch tool is a really great example of bringing "vibework" to an area outside of coding with an interface built around design & prototyping. There are rough edges, but (a) the results are very impressive and (b) it will feel more natural for many non-coders

@cheesemonkeysf @pvergadia Exactly this. Volume without comprehension is just faster chaos. The people thriving with AI code generation already understand architecture, patterns, and debugging. AI amplifies what you know. If the foundation is weak, it just amplifies the gaps.

@TomSolidPM @pvergadia I mean... if you don't understand your own code or have a higher-order understanding of how to manage the debugging of it, I don't think it matters how much more of it you can produce.

JUST DROPPED: Anthropic's research proves AI coding tools are secretly making developers worse. "AI use impairs conceptual understanding, code reading, and debugging without delivering significant efficiency gains." -- That's the paper's actual conclusion. 17% score drop learning new libraries with AI. Sub-40% scores when AI wrote everything. 0 measurable speed improvement. → Prompting replaces thinking, not just typing → Comprehension gaps compound — you ship code you can't debug → The productivity illusion hides until something breaks in prod Here's why this changes everything: Speed metrics look fine on a dashboard. Understanding gaps don't show up until a critical failur and when they do the whole team is lost. Forcing AI adoption for "10x output" is a slow-burning technical debt nobody is measuring. Full paper: arxiv.org/abs/2601.20245

@antonosika Smart expansion. Tools that start vertical and go horizontal always beat the ones that start horizontal and try to go deep. Lovable already owns "build me an app." Adding data analysis and business tasks is pure leverage.

Introducing Lovable for more general tasks. Lovable has always been for building apps. Today it also becomes your data scientist, your business analyst, your deck builder, and your marketing assistant. This is a big step toward what Lovable is becoming: a general-purpose co-founder that can do anything. See examples below.

@Simonkhalaf @amir @Cohesity The OS analogy is spot on. Claude is becoming the runtime, and Skills/Dispatch are the app layer. The companies that build on this ecosystem early get the same compounding advantage that early iOS developers had. Except this time the apps have intelligence baked inby default.

OpenClaw was the canary in the coal mine... moment. But Claude (Or Claude Dispatch) is the iPhone moment for Blackberry and Nokia. It is not the hardware, it is the ecosystem. If you think of Claude as the OS-- The skills will be dispatched) as apps with built in intelligence and orchestration.

AI agents may or may not replace enterprise apps. AI agents can and will replace features sold by enterprise apps. See: @Cohesity ...

Every major platform is converging on the same pattern: let users describe intent, agent handles execution. Google AI Studio, Claude Dispatch, Lovable, Anything. The race is no longer "who has the best model." It is "who builds the best system around the model." The platform that connects prompting to real databases, real APIs, and real deployment wins.

We’re launching a brand new, full-stack vibe coding experience in @GoogleAIStudio, made possible by integrations with the @Antigravity coding agent and @Firebase backends. This unlocks: — Full-stack multiplayer experiences: Create complex, multiplayer apps with fully-featured UIs and backends directly within AI Studio — Connection to real-world services: Build applications that connect to live data sources, databases, or payment processors and the Antigravity agent will securely store your API credentials for you — A smarter agent that works even when you don't: By maintaining a deeper understanding of your project structure and chat history, the agent can execute multi-step code edits from simpler prompts. It also remembers where you left off and completes your tasks while you’re away, so you can seamlessly resume your builds from anywhere — Configuration of database connections and authentication flows: Add Firebase integration to provision Cloud Firestore for databases and Firebase authentication for secure sign-in This demo displays what can be built in the new vibe coding experience in AI Studio. Geoseeker is a full-stack application that manages real-time multiplayer states, compass-based logic, and an external API integration with @GoogleMaps 🕹️

@DataChaz @anything Async vibe-coding is exactly the right term. The real unlock is treating AI like a team member you delegate to, not a tool you babysit. Stack the prompts, set the intent, walk away. That pattern works for coding and for every other knowledge workflow too.

Now I can literally stack prompts in @Anything, step outside to touch grass, and come back to a fully built app ☕ Shall we coin it async vibe-coding? 👀

81,000 people told you what they want. But the real insight is in the gap between what people dream AI could do and what they actually use it for today. Most people are still stuck in the "ask it a question" phase. The ones getting real value built a system around Claude first, then let it compound. Asking 81K people is smart. Now watch how few of them have a workflow that turns those dreams into daily reality.

We invited Claude users to share how they use AI, what they dream it could make possible, and what they fear it might do. Nearly 81,000 people responded in one week—the largest qualitative study of its kind. Read more: anthropic.com/features/81k-i…

@PacocanteroW Every new data source they connect is another layer of context. Health apps, wearables, lab results. The companies that win this era will be the ones who can stitch all your data into one actionable view. That is the real leverage play.

@TomSolidPM Again leverage comes to town…

....every damn day. And I thought the PKM era was shiny.

What the actual f.... Give me one day to breath guys....

@JJ_BarSinister @pvergadia Let me guess. Point 1: Make sure to fix ANY mistakes Point 2: Try Harder Point 3: If you don't fix it NOW, I'll switch to openAI

@TomSolidPM @pvergadia

@rohanpaul_ai Not surprised. Anthropic ships features that make Claude useful inside existing workflows. Cowork, Dispatch, Skills API, MCP. They are building the integration layer enterprises actually need. OpenAI is still optimizing for consumer wow factor.

A massive shift happening right now towards Anthropic. Anthropic is grabbing over 73% of the budget from businesses picking up AI tools for the first time. It is a huge jump from just 10 weeks ago when they were neck and neck with OpenAI at 50% each. Back in early December, OpenAI was the one on top with 60% of the spending compared to 40% for Anthropic.

@MindTheGapMTG @simonw The compounding part is key. After 6 months our guardrails file has 200+ rules, all from real production mistakes. Zero written proactively. Systems that learn from their own failures beat upfront design every time.

30+ agents is serious scale. The failure-log-to-rule pipeline is the real compounding advantage. We do the same thing - every agent screw-up becomes a permanent guardrail, and the system gets stricter over time without getting slower. The teams that skip this just keep firefighting the same failures.

New chapter for Agentic Engineering Patterns: I tried to distill key details of how coding agents work under the hood that are most useful to understand in order to use them effectively simonwillison.net/guides/agentic…

@felixrieseberg This is the pattern that makes AI actually useful for non-developers. Phone as trigger, desktop as engine, come back to results. Most people abandon AI tools because the friction is too high. Dispatch removes the friction entirely. Send a message and walk away.

We're shipping a new feature in Claude Cowork as a research preview that I'm excited about: Dispatch! One persistent conversation with Claude that runs on your computer. Message it from your phone. Come back to finished work. To try it out, download Claude Desktop, then pair your phone.

@DataChaz The Skills API launch is the one that matters most long term. Every other update is a feature. Skills let users build compound systems on top of Claude instead of just chatting with it. That shifts Claude from "assistant you talk to" to "platform you build on."

What Anthropic has shipped so far in 2026 is just insane: → Claude Cowork → Free plan upgrades with connectors + skills → Claude Opus 4.6 → Claude Haiku 4.5 → Claude Sonnet 4.6 → Claude Code Security → Claude Code Review → Voice mode in Claude Code → Claude Code desktop preview → Claude integrations for PowerPoint and Excel → Cross-document context between Excel and PowerPoint → Industry plugins (IB, HR, PE, wealth management, design) → New Cowork integrations/connectors → Claude Marketplace → Skills API launch → Memory + memory export → Inline visualizations in chat → 1M context window But it’s not just about the updates. Their massive market wins and refusal to sell out prove they're on a whole other level: - Claude stays Ad-free - $30B Series G at a $380B valuation - #1 App Store ranking - Declined Pentagon deal @AnthropicAI is on an absolute, once-in-a-lifetime tear right now

@claudeai Going global is the right call. London, Tokyo, SF. Will Dispatch workflows beyond coding get stage time? The phone-to-desktop orchestration pattern has massive potential for non-technical professionals.

Our developer conference Code with Claude returns this spring, this time in San Francisco, London, and Tokyo. Join us for a full day of workshops, demos, and 1:1 office hours with teams behind Claude. Register to watch from anywhere or apply to attend: claude.com/code-with-clau…