we live in mass indie times • brand strategist • european
San Francisco, CA
Joined June 2008
- Tweets2.2K
- Following422
- Followers200
- Likes14.8K
Pinned Tweet
I’m going to see a lot of friends in europe next week, so I’ve been shopping for SF souvenirs to bring. sharing my faves here:
1. open editions on valencia: great for art-inclined people. i got a sunprint kit there made at uc berkeley that feels incredibly special
2. aesthetic union in the mission: they do letterpress printing, so you’ll always find reasonably priced letterpress posters and prints by local artists and enthusiasts
3. song tea on sutter: amazing, high-quality teas
4. tartine: they have hats that are basically the premium version of the trader joe’s tote for any foodie, plus packaged cookies and biscuits that keep for a few weeks
alexa retweeted
i think what was actually happening during the Who's On First scene was that she simply wasn't paying attention, because a man was making her watch a youtube video
alexa retweeted
Bib Gourmand is the actual sweet spot. Excellent cooking with zero performance anxiety and no 23 course foam lecture. All the upside with none of the star chasing pretentious bs.
alexa retweeted
restaurant i went to today had a 17% service charge included, which made me want to add another 3% to bring it to 20%, except tipping 3% felt insulting, so i added 10% instead. on top of the bill that already included the 17%. so i ended up tipping 28.7%. incredible psychological warfare
alexa retweeted
I built a site to map silicon valley’s hwy 101 billboards and track the lore (80% are ai & tech).
one founder told me he first saw his board in the wild here.
now each ad has its own link, easy to share, like this OpenAI @steipete feature: 101ads.org/b/c8496954/
alexa retweeted
they said the devil was in the details. we said, why? why should the devil be confined to the details? why should access to the devil be limited to a privileged few?
today, after years of research, hundreds of millions of dollars in funding, and countless sleepless nights, we’re proud to announce that we’ve successfully removed the devil from the details.
he’s everywhere now.
this is what democratizing access to the devil looks like. and we’re just getting started.
we’re hiring!
garden tomatoes are tomato slop. you only like them because of their unparalleled flavor and their ability to transport you back to a fleeting moment in your grandmother’s garden
sometimes I can’t stand rereading my own prompts because all the typos make me cringe
I find it ironic that “logotype” (from the greek “logos”, meaning “word” + “typos”, meaning “mark”) was shortened to “logo” which literally just means “word”. we turned words into visual symbols only to end up calling those symbols words again
me: it’s gonna rain next week
men: yes but have you seen the exact shape of the storm that’s heading your way
sucks to be banned from waymo at 19 and then be unable to take one for the rest of your life because there are no more taxis or ubers
This Post is from a suspended account. Learn more
alexa retweeted
the IoT debt is going to start showing up soon. when labs people predict catastrophic cybersec attacks, the mean this. the US Cybertrust Mark, sortof defunct, can't do anything for this - medium.com/@danberte/2026-th… (Jan 26)
So Anthropic claimed it ran a 26B AI model across fifteen $55 Android TV boxes.
Setup
📺 15× off-brand Android TV boxes
💾 4GB RAM each
💰 ~$55 each
🧠 Gemma4-26B-A4B
⚡ ~0.39 tok/s
Total hardware cost = ~$825.
The model was sharded across the boxes using pipeline parallelism, so each tiny machine only had to hold and execute part of the network.
Yes, 0.39 tps is sloooooooow.
But, that 26B model was actually running across a pile of cheap consumer boxes that were never designed to be AI servers.
🎯And Anthropic also said that Qwen3.5-9B agents managed to set up inference successfully on 8 of 9 hardware types they tested, including Macs, Jetsons, and RTX systems.
Source: Anthropic Frontier Red Team
⚠️Note that the 15-box demo was distributed inference. The self-replication experiments in the paper were separate.
ALT https://www.anthropic.com/research/self-replicating-ai
