CEO of Stateshift. GTM for technical people who hate being marketed at. 📓 Author ‘People Powered’ by Harper Collins. Wrote 'People Powered' (@harpercollins).
Bay Area, California
Joined May 2008
- Tweets29.4K
- Following6K
- Followers30.9K
- Likes11K
How a leaked file started the Open Source AI revolution... 👇
ft. @Meta, @llama, @WizardLM_AI, @ggerganov, @finkd, @StateshiftHQ, @MistralAI, @Raspberry_Pi
On a flight and literally can't tell whether the smell is from the food or whether someone threw up. Not even joking...
Heading out to @AllThingsOpen, @AgenticAIFdn @linuxfoundation AGNTCon, and @OReillyMedia FooCamp. Hope to see some of you in person!
Tomorrow at 4.55am I will be waking up to pre-order an iPhone Duo...Apple's first foldable. Am I a total numpty for wanting one, or are you excited for a folding iPhone too?
Image Credit: Tom's Guide.
TODAY at 10am Pacific: a FREE hands-on workshop for how to create your first 10 videos for YouTube and social media - creating ideas, scripting, editing, thumbnails/titles, which metrics to track and improve. Register here: 👇
linkedin.com/events/75081949…
Jono Bacon retweeted
GitHub stars are a vanity metric. Before you raise the pitchforks, here us out.
They feel great, photograph well, but tell you almost nothing about whether anyone actually uses your project. A star is basically an applause, but not actually product adoption.
We've worked with dozens of dev-focused companies and the ones with genuinely healthy OSS projects have stopped putting star counts on the slide. They track three behaviors instead:
1. Package downloads and active installs (npm, PyPI, Docker pulls). Someone pulled the code onto a machine. That's a decision, not a bookmark.
2. Issue signal-to-noise. Repeat contributors filing real bugs = people running you in production and hitting the sharp edges. Drive-by reactions mean nothing.
3. Time-to-second-PR. First PRs are curiosity, second PRs are commitment. In the healthy projects we've measured, a returning contributor shows up inside 30 days. In the dead ones, they never do.
Stars tell you people noticed, but these three will tell you they really cared.
+1
Developers are NOT impulse buyers.
You're not selling ketchup on a shelf. You're selling a tool someone will probably use everyday.
Flashy campaigns get attention but they don't earn trust. Devs read the docs. Test the workflow. Ask their peers. Hunt for proof the thing actually solves a problem they care about.
So stop trying to impress them. Help them make a better decision.
That means plain language, real examples, technical depth, and proof you get the problem they're in.
Be useful before you ask for anything.
How we built @ubuntu's 65M+ community (and what I'd never do again)
ft. @sabdfl, @mdz, @Canonical, @castrojo, @mhall119, @makoshark and many others.
Chart of my YouTube channel's growth...and an illustration of "less is more."
Like a lot of founders....I decided on a cadence of content - 2 videos a week w/ shorts every day. My stupid theory: more content == more growth.
Well, the result was rushed, uninteresting, boring content.
So...a month ago we made the decision to scrap the schedule and focus on every video counting. Now we ship fewer videos, but each performs waaay better, because we are super-intentional about what videos to greenlight...we spend time discussing and fine tuning the edit/thumbnails/...and the results are better but also the process *feels* better because it is anchored to quality.
Are the videos perfect? No, they never will be and we have a lot to learn, but I think we are on the right path.
So, do less...and make the less you do really count. 🤘
