Tom Green

@haydenbleasel Thanks for all you've done for the community. So many cool projects making up so much of my stack. Looking forward to seeing what you're moving on to. Good luck !!

Today is my last day at Vercel. I've had an absolute blast over the last year. I got to work on a lot of my own projects that helped form the basis for our AI Cloud, helped other teams ship their great ideas, and worked alongside some of the most talented people I've ever met. I'll have news on my next adventure for you on Tuesday but for now, wanted to reflect on my time here and the projects I worked on. Read on if you're interested ↓

@republicanolat @d_p_ball @Microinteracti1 @LindseyGrahamSC @POTUS Provide proof that there was no proof....right

@d_p_ball @Microinteracti1 @LindseyGrahamSC @POTUS He should provide proof. Saying something without confirming documents is just hearsay. Being that he is a counter terrorism director, surely he has evidence.

Just spoke to @POTUS about our European allies’ unwillingness to provide assets to keep the Strait of Hormuz functioning, which benefits Europe far more than America. I have never heard him so angry in my life. I share that anger given what’s at stake. The arrogance of our allies to suggest that Iran with a nuclear weapon is of little concern and that military action to stop the ayatollah from acquiring a nuclear bomb is our problem not theirs is beyond offensive. The European approach to containing the ayatollah’s nuclear ambitions have proven to be a miserable failure. The repercussions of providing little assistance to keep the Strait of Hormuz functioning are going to be wide and deep for Europe and America. I consider myself very forward-leaning on supporting alliances, however at a time of real testing like this, it makes me second guess the value of these alliances. I am certain I am not the only senator who feels this way.

@Shpigford @RumoredAI Congrats! Love the ballsy pricing structure...though I am priced out 🤣

i've been heads down lately working on a new thing: @RumoredAI and today it's available! everyone's familiar with SEO and everyone's becoming more familiar with AEO/GEO (which is optimization for AI). yes it's interesting to know what terms/phrases surface your business, but what nobody has tackled is what to do when AI is getting your business *wrong*. and we found that AI hallucinates business facts for quite literally every brand. rumored.ai surfaces what AI is saying about your brand, what it's getting wrong, how you compare to your competitors and (most importantly) the exact things to do to fix those issues. you get a ridiculously in-depth interactive threat report covering 12 sections: from executive summary and active threats to competitive analysis, schema audit, and a prioritized action plan with copy-paste fix prompts. this isn't a subscription (yet?). it's a one-time purchase of an in-depth audit of your business. launch price is $25. but the price goes up by $25 each time someone purchases. 📈 have been testing this with a lot of companies and the response has nearly universally been 🤯. i think you'll love it.

@hamza_automates 500

I BUILT A $5,000 AI AUTOMATION FOR ROOFING COMPANIES And you can copy it. This one handles everything roofers hate touching: • Reads the incoming job requests • Pulls out the project details • Checks availability • Builds the estimate using their exact pricing logic • Sends the proposal • Books the job • Follows up • Invoices • Collects payment • Requests the review All without the owner unlocking their phone. Same CRM. Same calendar. Same inbox. Just a system doing the work instead of a human. If you want the breakdown (how it works, what tools run it, and the logic behind the whole thing): Comment “500” and I’ll send it to you. (must follow for DM)

@Rockabye2512 @DerekPederson3 Great summary of a shit situation. Thx

@DerekPederson3 mining the shit out of the waters. Minelayers are not big. Minelayers are, however, many. This means they will swarm a strait so narrow you can shine a flashlight on one side and see the light from the other. And mine the shit out of it. They will also send drone (cont)

So the outcome so far is: - the regime is weakened military but hardened politically - Iran destroyed their relationship with several countries that had been neutral in the Middle Eastern Cold War like Qatar, Oman, and Turkey that will now likely lean towards the anti-Iran coalition in the future - Iran has lost a lot of influence in both Lebanon and Iraq (Maliki is out, looks like Sudani will get another term as PM) and in Lebanon it might be for good if Hezbollah can be disarmed with the help of Syria and Israel - No obvious off-ramp to the war as the IR has rejected a ceasefire but we might be able to keep the Straight of Hormuz open via tactical operations - The opposition remains too disorganized relative to the hardliners in the government to overthrow the regime, despite a majority of Iranians desiring this outcome and even factions of the government probably wanting to break with the Mullahs - The only thing the US could do to *guarantee* such an outcome would be a ground invasion, for which there is no support in the US Does this all sound correct?

Anyone using Claude/got a Claude skill for UK accounting yet ?? 👀

@ivanburazin @cmwalker Hard agree. There should be an opt out setting so people can't send you voice notes. Go transcribe it yourself if you can't be arsed to write it. Don't make me do it for you

Voice messages should die. I hate everyone who sends them. When someone sends me a voice note, I don't look at it for days. It's the most selfish form of communication. You're offloading your typing time onto my listening time. You get convenience and I get locked into your pace. Can neither skim thru nor reference them later.

@jordan_ross_8F Link

I charge $25,000+ to build these automations for agencies. Here are the first 10 we install every single time — for free. We've worked with 1,250+ marketing agencies. Our clients have added $500M in annual recurring revenue. Our 20 engineers log over 6,000 hours a month solving exactly one problem: agencies bleeding time on tasks a $20/month tool could handle. Most agency owners want to automate. They just don't know where to start — so they don't start. Meanwhile, their team is burning 20+ hours a week on work that shouldn't require a human. One example: data analysis alone costs the average agency 1 hour per client, per week. Multiply that across 15 clients and you've got two full workdays gone before anyone's opened a brief. I put together a free Doc called The 10 Agency Tasks We Automate First. Each task includes: → What the task is and what it currently replaces → How long the setup actually takes (most are under a session) → The exact tools behind each automation with real pricing ($20–$800/month depending on stack) → Why we use a 30-minute threshold to decide if something's worth automating → A prioritization framework so you know where to start with your specific agency → Entry-level cost breakdown — first five automations can run ~$100/month on tools you likely already have → The 6 tools our engineers rely on across these 10 automations (including the MCP solution we built — the first of its kind for agencies) → A clear view of what 20+ hours of recovered capacity actually looks like across a team You don't need a dev team. You don't need a big budget. You need the right list in the right order. Comment LINK and I'll send it directly. If we're not connected, add me first so it reaches you.

@0xhashlol What are the benefits ?

@tomgreeEn The CLI learning curve is brutal but worth it long-term. Context switching between terminal/editor actually makes you think more deliberately about what you're doing. That said, Cursor's autocomplete is addictive - hard choice!

Finally switched to Claude Code from Cursor today Are people really claiming that working in the command line is more productive than in a fully integrated IDE?? I'm sure I'm at the bottom of a learning curve but holy shit this stuff is painful compared to the alternative

@Shpigford Not all heroes wear capes

@Shpigford @openclaw This is amazing - can you turn it into a skill??

another 🦞 use case: one of our cars got sideswiped and now needs some repair work done. there are dozens of places within a few miles of us, but figuring out which one to go to for an estimate required some effort. effort that i just offloaded entirely to hank (my @openclaw instance). things hank did: * google searched for the top rated repair shops closest to our house * called all of them to get lead times on repair work and find out if an appointment is required for an estimate * ranked them by lead times + distance + ratings * saved all the info in a notion doc (location, contact info, ratings, lead time, appointment requirements) all i did was give it about 3 sentences of instructions and then i went and did other things while it chewed through the job over the course of about 10 minutes. the calls alone were a massive time-save as it did them in parallel. so. freaking. cool.

@pbteja1998 Great write up. Do you think it's important that it's running on a dedicated machine ? Or is the level of self sufficiency that is the magical part ?

@michelle511 @True_McLofty @pepel_klaasa and...? L

@True_McLofty @pepel_klaasa Mainstream media

The U.S. Secretary of Commerce, Lutnick, behaved so outrageously at a dinner with his European counterparts in Davos that he was booed, and some – the head of the European Central Bank, for instance – actually stood up and walked out. Reportedly, Lutnick even received a personalized “booo!” from former U.S. Vice President Al Gore. 🍿 Looks like this whole realpolitik thing is starting to come back and bite everyone in a very tangible way

@haydenbleasel @mintlify @ollama @supabase @triggerdotdev @mastra @Cloudflare @elevenlabs @upstash @langfuse Great library, great update, thanks!! What're your thoughts/experience on letting the LLM know what syntax (mermaid etc) the renderer supports? Do you have any boilerplate you include in the system prompt? Would love an .md baked into the package that could be included

I've had a pretty amazing 6 months of building Streamdown. What started as my react-markdown wrapper for streaming is now being used to power AI experiences at companies I love, like @mintlify @ollama @supabase @triggerdotdev @mastra @Cloudflare @elevenlabs @upstash @langfuse, @dify_ai and of course here at @vercel. Today, we're releasing the first major version update. More details below, keen to hear what you think! 👇

@samuelekpe Brilliant thread, thank you 🙏 I would add one more: Logic > LLM based decision making It will almost never make sense to let an agent decide something that could be determined by code. Code= faster, cheaper and deterministic. Optimise workflows by maximising use of logic Vs LLM

We’re entering the age of Autonomous Organizations. The real moat is understanding that AI agents need management, culture, and process; exactly like people. Here's a thread with key insights from 12 months of building the OS for fully autonomous workforce.

@zobeir yes pls !

built an AI that simulates how Apple will review your app so you know if it'll get rejected before you submit

@daniel_nguyenx @JamesIvings Amen. Also did this once as felt like effortless redundancy, got burnt and never again

@JamesIvings I did this like 10 years ago and it f’ed up and I ended up losing data lmao

someone just told me they backup their .git directories with dropbox and now I feel violated

@VibeMarketer_ Script

the best way to find ad ideas, is to scroll TikTok, but nobody wants to do that for hours on end... so I built an AI agent that can do it for you, completely hands-off just give it a keyword and it: – scrapes the top-performing TikToks – pulls the pain points, trigger events, and emotional hooks – rewrites the viral script for my product or service want the full n8n flow + set up guide? like + comment “script” and I’ll send it over (must be following so I can DM)

@josephpalbanese V cool testflight pls !!

new project: cleo (kindle + llm) 📚 cleo is an ios app that pairs whatever book you're reading on kindle with an llm (o3) — ask questions, get recaps/summaries, or listen/discuss (think audiobooks but interactive). the llm has context on exactly where you are + the book contents. best part? no complex setup needed. just link your kindle account and that's it. it just works. reply / rt for testflight invite 👇