Hey! Tal here. Maxim and I created Familiar to capture our screen (and clipboard) every few seconds and save it as markdown. That way our local agent can use that as context (through a cron, skill, or slash command).
Our early testers are showing us how to use it:
- Update Claude's skills/memory based on their workday (in a scheduled task/heartbeat)
- Typing "help me with what I'm working on right now" without having to prompt/describe
- Enriching meeting transcripts with what was actually on screen (and vice versa)
- Forking it into their coaching app so coaches can see what learners did between sessions
- Someone new to tech used Familiar during a trial week at a YC startup, so that AI could coach him every few hours (and got the job)
We've consistently seen AI use Familiar context like a "router" layer. Recently, my agent "saw" that I spent a long time on a document, so it fetched the full doc directly. We've also seen the agent traverse the markdown, then decide to fetch the original image (so cool).
Familiar uses Apple's native OCR, deletes screenshot images after 48 hours, and redacts passwords/credit card numbers/SSNs/API tokens/etc. We'd love contributions on what else to block: https://github.com/familiar-software/familiar/tree/main/src/... or in general ways to improve privacy.
We stand on the shoulders of giants: screenpipe, rewind, dayflow, etc. Since then: 1) Local agents got good at handling massive amounts of messy text files 2) Local agents have their own memory and skills systems
Familiar is our "bitter lesson" version: just hand over context and get out of the way. The right way to do that piece is open source/free/offline.
This coheres with my dating experiences as well. The medium may have changed, but we're all still people, with mostly the same spectrum of motivations we've always had.
My issue with dating apps is that they're too efficient. Our attractiveness fluctuates dramatically based on context. If we're doing something we love, doing something interesting, or exhibiting a particular behavior or trait, we can be far more attractive than any five curated photos will show. And vice versa.
Dating apps remove all contextual variety and flatten us out. And in that case, "efficiency" is actually inefficient.
> The "flow" that imaginative people love so much has a darker cousin that prevents you from pausing to savor life...
I'm glad he pointed out this seemingly small detail. This took me a very long time to understand.
EDIT: It reminds me of another great post by Paul Buccheit. It's so important to have the 'heroes' of startup culture explicitly spell out these values:
> I worry that perhaps I'm communicating the wrong priorities. Investing money, creating new products, and all the other things we do are wonderful games and can be a lot of fun, but it's important to remember that it's all just a game. What's most important is that we are good too each other, and ourselves. If we "win", but have failed to do that, then we have lost. Winning is nothing.
I understand why you might arrive at that conclusion, but think about another level: What did Jessica Livingston lose or miss out on by not promoting herself to the public?
I would argue that she missed out on nothing that matters.
I think that's the calculus that she figured on when she made the decision not to be a public figure.
I suspect that the recent articles playing up her role are because she & YC realized that there is something they're missing out on: millions of potential female founders are not starting companies because they don't have strong role models. This hurts YC in a financial sense, but even beyond that: if your mission is to increase the amount of startups and innovation in the world, then having half the world's population disqualify themselves because they don't have many good examples of it being done before is losing out.
> In fact, the real moral of the story is that the tech press is almost entirely bullshit, and that is mostly to be avoided.
Any interaction I've had with the tech press from the "written about" side of things (or reading about close friends) has given me less and less faith in what I read about those I don't know.
I'd venture that's valid for any interaction with journalism these days. You read a story and if you are part of the story you realize how wrong and/or skewed the whole thing really is. Then you start to question how many other stories or articles are the same. Eventually you realize you can only take them with a grain of salt.
I was close to deleting my Facebook account when I discovered the News Feed Eradicator[1]. This way I can still use messages, groups, events, etc. without having to experience the dreaded news feed. It really improved my quality of life.
Another major step I took was to stop posting. That way I had no expectations of feedback and logged in far less in general.
I just started using this, and it's wonderful. I previously was using Stayfocusd to limit Facebook time to 5 minutes per day, which I depleted quickly. Sometimes I got cut off in the middle of writing a reply to someone, or even doing something for work. Now I don't have to restrict myself anymore, since there's nothing to mindlessly scroll through.
Reading the feed was like playing slots. It pays out just enough to keep you playing, but on net you're still losing. "Maybe that next post will be interesting! Oh, nope, just another story about Donald Trump. On to the next post..." Never again.
Thanks for introducing me to this extension. I've been using my hand to cover news feeds on fb, twitter, etc when I need to go in and respond to a message or something. Although it's easy for me to do that as social news feeds don't hog my attention as much as they induce anxiety.
Our early testers are showing us how to use it:
- Update Claude's skills/memory based on their workday (in a scheduled task/heartbeat)
- Typing "help me with what I'm working on right now" without having to prompt/describe
- Enriching meeting transcripts with what was actually on screen (and vice versa)
- Forking it into their coaching app so coaches can see what learners did between sessions
- Someone new to tech used Familiar during a trial week at a YC startup, so that AI could coach him every few hours (and got the job)
We've consistently seen AI use Familiar context like a "router" layer. Recently, my agent "saw" that I spent a long time on a document, so it fetched the full doc directly. We've also seen the agent traverse the markdown, then decide to fetch the original image (so cool).
Familiar uses Apple's native OCR, deletes screenshot images after 48 hours, and redacts passwords/credit card numbers/SSNs/API tokens/etc. We'd love contributions on what else to block: https://github.com/familiar-software/familiar/tree/main/src/... or in general ways to improve privacy.
We stand on the shoulders of giants: screenpipe, rewind, dayflow, etc. Since then: 1) Local agents got good at handling massive amounts of messy text files 2) Local agents have their own memory and skills systems
Familiar is our "bitter lesson" version: just hand over context and get out of the way. The right way to do that piece is open source/free/offline.
reply