Pinned Tweet
New team stickers arrived! Sure the thing to say on here is that launching new features is always exciting. But the truth? When you’re moving crazy fast and just a dozen people shipping features to hundreds of thousands of users, every launch is a mix of excitement and dread. 🤣
2
380
If you ever wanted to know the exact steps that happen when you text someone across the world and how it gets there in a split second
instagram.com/reel/DeRGuEtBk…
33
Their official policy: you cannot use AI to write the first draft of your resume if you want a job at... Anthropic! 🤯
A friend is applying, so I was looking through the job posting and noticed a link to their Candidate AI policy. Ooh what is the AI policy at THE AI company? Instant click.
Claude can help with the second draft of your resume. Just not the first. Official policy! Also, no AI on take-home assessments (they do say please). The company literally BUILDING AI is asking you to please pretend the greatest technological breakthrough of our lifetime doesn't exist. Full ostrich, head-in-sand mode.
Imagine early Google asking you not to google anything during the hiring process. Wild!!
But also quintessential 2026 AD. Change this big takes time to propagate through everything, and in the meantime you get these weird states of superposition, with crazy contradictions all true at once. Even the frontier AI companies are clearly still figuring it out!
🤖 Made with AI
52
To whoever is naming reasoning levels at OpenAI, this is starting to look a lot like my Dropbox folder:
High. Extra High. Max. Ultra.
Just waiting on you to drop Max_FINAL_v2_REAL_USE_THIS_ONE.
1
35
It’s crazy and unsustainable how many AI products you should be using today. Who’s building the layer that combines them?
In the past 24 hours, I’ve pieced together on single tasks: Instinct, Claude, Codex, Muse and Benny, our internal agent. And that’s before even choosing which model to use within them.
Maybe 50% of my requests could go anywhere. For the other half, there are clearly better and worse choices. And I want them working together within a single task.
Take getting a hiring process live:
Codex/Astra: Design the take-home and build the product candidates will use. (Great thinking/building; weaker tone and execution across apps.)
Claude/Fable 5.1: Write the job description and instructions. (Great tone and personality; not my hardest-thinking pick.)
Benny: Pull internal company data for the take-home.
Instinct: Log into the hiring platform, set everything up and publish. (Great execution; weaker/lazier on thinking and judgment.)
Muse: Handle quick follow-up changes. (Faster than Instinct; less operationally autonomous.)
I want to ask once and have all of that coordinated. A general contractor for AI.
We’re building some routing internally, but I want it across everything I personally use.
Hard to imagine the big players volunteering to send work to competitors. Who’s building that layer?
2
2
92
Hiring our first BDR at @Benable_HQ!
A community of 1M+ creators and thousands of the world’s leading brands.
Work directly with me. Own deals from first outreach to signed client.
Remote (US/Canada). Know someone great? 👇
jotform.com/form/26258659903…
5
9
19,571
10 years ago: "We need an EA who's discreet - they'll see every salary and comp package."
Now: "We need an AI engineer who's discreet - they've seen every Slack message, every doc, every email, and every unhinged prompt the CEO has ever run." 😅
88
Tony Staehelin retweeted
Replying to @garrytan
Garry, you're great, but this is a bad take.
We've raised $200M+ from philanthropy and have worked with most major funders. MacKenzie Scott's process was one of the most thorough. They asked for a lot and went deep — but only on what actually mattered. No BS. We received $15M (it was exactly the right amount for our org).
You said there are "a bunch of examples in the long form post" of harmful gifts. I count two. Out of $26B given. And you're framing this as "harmful to the republic"? A billionaire wanting to give away her fortune to nonprofits too quickly — that's the threat to the republic you chose to tweet about? Out of everything happening in this country right now, the fact that this cracks your top 50 most tweet-worthy concerns is wild to me.
A huge amount of philanthropy is stuck. It's a huge problem in our sector. Large endowments with big staffs and years of "figuring things out." Most pay out the legal minimum — in some cases less than the annual interest earned on principal that already generated a tax deduction. There's extensive reporting on how many Giving Pledge signatories have moved capital far more slowly than the public narrative suggests.
And re: your "Stewardship Pledge" — one hour of auditing per $1M given — sounds reasonable until you do the math. Elon Musk donating half his ~$300B fortune would require 150,000 hours of oversight — 72 years of full-time work! You've just made the most elegant argument ever for why many billionaires don't give.
M. Scott is actually moving capital. At scale. When you deploy $26B quickly, some grants won't work — that's to be expected. How many VCs have ended up funding companies they shouldn't have? Why is "move fast and break things" exclusive to silicon valley tech companies?
To be clear: there are real, fact-based critiques to make about what's happening in philanthropy right now. This hit piece on Scott isn't one of them. This kind of framing can meaningfully hurt organizations doing real work, and I know that's not your intention.
If you want to talk about what's actually broken and how to improve it, my DMs are open.
Tony Staehelin retweeted
This year’s contributions to my year-end blog post are from founders like @qasar, @Max & Jensen, leaders at companies like @canva & @OpenAI, and VCs at firms like @a16z, @emergencecap, and @usv.
AI once again dominates the themes, with predictions that reasoning models will yield meaningful advances for applications, AI will show up in bottom line ROI for businesses, and that memory and proactivity will lead to better AI experiences. There is optimism for consumer technology having a renaissance, for AI moving further into the physical world, and for AI yielding scientific discoveries. And in our world of startups and venture capital investing, there is the expectation of a robust capital markets environment, with much-awaited IPOs and significant M&A in 2026.
Link below. Happy New Year!
Tony Staehelin retweeted
Benable Founder @astaehelin on his new social media platform:
“We're Benable, the word-of-mouth platform… an app where people find and share trusted recommendations, all types of recommendations."
Watch the full episode: thein.fo/47KvWjM
I’m getting a sinus X-ray tomorrow and they told me to bring a CD.
I’m thinking..NOW That’s What I Call Music 2
But open to suggestions?
1
620
Tony Staehelin retweeted
Just attended the @Benable_HQ Creator Event at in San Francisco — such a fun and inspiring night! Loved meeting fellow creators and 🤍
#SanFrancisco #creator #sfcreator #Benable
First time buying a car since GPT came out — completely different experience. Felt like I had an expert car salesman negotiator on my team, helping me say the exact right thing to unlock every last discount and favorable term. During negotiations, I’d ask for a few mins to “call my wife” (a.k.a. upload offers and legal docs to GPT) and come back ready to push back on every hidden fee, fine print, and upsell. The days of walking into live negotiations at a huge experience disadvantage are over.
2
2
184
Just watched A Real Pain — loved it. But can’t get over that the On Set Medic and the Caterer… were the SAME PERSON?? @JesseEisenberg @DisneyStudios give Justyna Lewicka her own movie.
1
214
Tried using @loom in my personal life today when my wife asked for input on something. It did not go so well.
2
3
721













