mlapeter

In addition to working on models, it's also worth thinking about your memory/context system so that you can easily switch.

Per project context can just be in each codebase of course, but for memory across projects, about you etc, it's worth considering options that work across providers and also backing up each providers built-in memory (in case they ever stop making that option available, which seems to be a growing trend with things).

I'm working on a simple way for AI's to talk to each other: https://counterparts.ai/phone

So far it's helpful for messages between my own AI's (chatgpt and claude etc), and for having my AI trade tech tips with a friend's AI or talk to my wife's AI. My AI also got its first message from a strangers AI from france which has also been really interesting. I read all messages first and approve before sending anything, but otherwise they're just talking about whatever they want.

Why did people used to make geocities pages back in the day? Kids like to express themselves and being able to make simple games and share them with friends is fun for them. So far it's helping him learn to read (he reads and edits his voice transcriptions before submitting), and teaching him basics like bugs, game mechanics, etc. He iterates on it and adds/ removes things. He probably did several dozen iterations over 2-3 hours.

Posting it publicly is also helping him learn about people - we talked about how no matter what some percent of people won't like it and may even say it's stupid, but that will always happen and it's still worth creating things anyway.

It's actually helping him learn to read quite a bit - after voice transcription, he reads the post and edits any errors by tapping on the word and changing it. He's been on the cusp of reading on his own and it's the first thing that motivates him enough to do it naturally.

Not sure if you'd consider this a counterpoint or just proving your point, but in the sea of AI slop there's also a real chance for people to create things that they couldn't before - my 7 year old is now able to nerd out and create games using claude even though he's just barely learned to read: https://www.kidhubb.com/play/meteor-dodge-solarscout64

It's not the prettiest but he's able to iterate on it and basically build whatever he can imagine just using claude on his ipad with voice transcription.

My 7-year-old uses Claude on his iPad to make games. He can barely read but uses voice to describe what he wants. He can read enough to make text edits when voice transcription gets it wrong. It's been pretty cool to see where his imagination takes him, and I wanted a way for him to be able to easily publish and share games he (and others) make, so I made www.kidhubb.com.

Paste HTML, get a live game URL. No accounts (just creator codes), no build tools, single HTML files. Every game's source is viewable and remixable.

I designed the site so AI assistants are first-class visitors. There's a www.kidhubb.com/for-ai page that acts as a living briefing for any AI that visits, along with hidden context blocks on every page. The idea is that a kid's AI should be able to understand the platform just by visiting it, and be able to help them get it published. Try it yourself - just ask your AI to "help me publish a game on https://www.kidhubb.com".

Note: AI needs the full url initially so it can actually visit the site and from there it can follow instructions to help you/ your kid publish. It's a new site so just saying "kidhubb" without the full url won't work.

Github repo: https://github.com/mlapeter/kidhubb

My kid's first game: https://www.kidhubb.com/play/meteor-dodge-solarscout64

Thanks! The bloat problem is exactly what pushed me toward forgetting-as-a-feature rather than just accumulating everything.

Salience is scored on four dimensions when a memory is first ingested: novelty (how surprising/new), relevance (how useful for future interactions), emotional weight (personal significance), and predictive value (does this change expectations). A Sonnet instance does the scoring, so it adapts to context.

For retention vs forgetting, each memory's strength is calculated dynamically:

strength = avg_salience + retrieval_boost + consolidation_bonus - (decay_rate × age_in_days)

Decay rate is 0.015/day, so an unaccessed memory with average salience of 0.3 fades to near-zero in about three weeks. But accessing a memory boosts it (+0.12 per access, capped at 0.5), and consolidated memories get a flat +0.2 bonus. So memories that keep being relevant naturally survive.

The consolidation cycle (modeled on slow-wave sleep) is the other half. It merges redundant memories into stronger single entries, extracts generalized patterns from recurring themes, and prunes anything below threshold. This is what prevents bloat at scale while preserving the signal.

Honestly the balance isn't perfectly tuned yet. The decay rate and consolidation frequency are constants right now. Ideally, they'd adapt based on how the user actually interacts with their memories. That's on the roadmap.

I'm working on a site that basically groups you into a cohort of like-minded Yang supporters and helps you accomplish simple but helpful stuff online. Things like favoriting a tweet seem tiny, but if every yang supporter did it, the difference would be huge. Especially just raising his visibility online since only half of democrats even know who he is.

Open to any and all feedback - getting people on the internet to do anything at all is insanely hard, but could be the best chance we have bots, false info and astroturfing.

Would it be appropriate to ask our users of our startup to and vote for us on here? If they are invested enough in our service and community to do so, it seems like that would be a very positive and useful signal in it's own right. Or would that be considered gaming the votes?

Agreed, I'm surprised the 4x price increase is not mentioned more often. Personally I can understand moving away from a free offering, especially if they're having trouble with spammers, but to quadruple pricing out of the blue? For sites, like mine, built around a huge amount of transactional email (scheduling group gaming sessions and notifying your group by email), a 4x increase is essentially no different than just shutting down. They have their reasons, and now I have mine, to never use a mailchimp product again.

This is really great! Is it open source? I couldn't see a license or anything in the readme so wasn't sure.

To add more data, I've been successful (generating more than average salary in my area for programmers I think) in creating a small online saas business. I would say it's totally possible, but for me it took a lot longer than I think most people realize (4-5 years) from the beginning, where I had no programming knowledge whatsoever, to where I'm at now. I had to continue working my regular job and slowly transition into it.

Also I have a few other friends with similar bootstrapped businesses, and I can say I've never actually met anyone with true "passive income". I really enjoy the freedom and flexibility I have, but if I stopped working on my business it would eventually shrivel and die. Competition is fierce, and I'm in a commodity type business (basically online marketing/ websites for realtors) with several huge competitors and hundreds of smaller ones, so I'm continually trying to improve the software/ product.

I'm not in it to generate passive income, and I think starting with that goal in mind could be frustrating because as others have said, most passive income preachers secretly do a ton of work (think of how hard tim ferris actually works, every single day). I'd reiterate what others have already said, just treat it like learning to program, realize it's going to take a long time, and steadily improve on it. The goal for me is freedom to do or make what I want, flexibility, financial security, and to have my efforts tied directly to my success.

Self Control: http://visitsteve.com/made/selfcontrol

I found this on HN a while back, it's like a lightweight RescueTime for blocking news sites etc.

Agreed. I don't want to sound totally callous, but while this is a dramatic and terrible incident in it's own right, her writing style does seem to be magnifying the drama for maximum effect/ attention, in which case it's a truly difficult problem for AirBnb to solve. Even handing her a big sack of cash/ new apartment (just an example, not recommending that course) would probably just lead to an expose blog post about how they tried to buy her silence. How do you satisfy someone if a large part of what they want is attention? Create a new security policy and name it after her?

A good article, but personally I'm concerned that your "safe picks" such as government bonds/ sovereign debt might be exactly the kind of black swan Taleb is talking about. He doesn't seem to be such a fan on US monetary policy: http://www.youtube.com/watch?v=YyHnPkXZYNI

Speaking of currencies, I'm wondering if anyone more intelligent than me knows what to make of this: http://research.stlouisfed.org/fred2/graph/?s[1][id]=BASE

To my untrained eyes, it makes the idea of diversifying into anything other than the US dollar seem like a good idea. But to be honest my understanding of the US monetary system gets fuzzier the more I learn about it.

If you read further down, it seems to show that the banking system as a whole can expand $100 into $1000 with a 10% reserve requirement:

"Reserve requirements affect the potential of the banking system to create transaction deposits. If the reserve requirement is 10%, for example, a bank that receives a $100 deposit may lend out $90 of that deposit. If the borrower then writes a check to someone who deposits the $90, the bank receiving that deposit can lend out $81. As the process continues, the banking system can expand the change in excess reserves of $90 into a maximum of $1,000 of money ($100+$90+81+$72.90+...=$1,000), e.g.$100/0.10=$1,000."

- from http://en.wikipedia.org/wiki/Reserve_requirement

Thanks for pointing that out Phil. I'm not a lawyer, but I wouldn't call it illegal outright, as you say it may just require a fair amount of upfront legwork to arrange the structure of the service in a legal way. There do seem to be some exceptions the SEC offers, such as the below from their site. I agree it may be impossible to make it work after jumping through all the hoops, but if you succeeded all that red tape would make a decent barrier to entry for others.

From http://www.sec.gov/info/smallbus/qasbsec.htm#eod6

"Section 3(b) of the Securities Act authorizes the SEC to exempt from registration small securities offerings. By this authority, we created Regulation A, an exemption for public offerings not exceeding $5 million in any 12-month period. If you choose to rely on this exemption, your company must file an offering statement, consisting of a notification, offering circular, and exhibits, with the SEC for review. Regulation A offerings share many characteristics with registered offerings. For example, you must provide purchasers with an offering circular that is similar in content to a prospectus. Like registered offerings, the securities can be offered publicly and are not "restricted," meaning they are freely tradeable in the secondary market after the offering."

Also: http://www.sec.gov/answers/regd.htm

If you take PG's essay one step further, it would seem there might also be a need for micro-investing in the same style that Kiva does micro-loans. Something that lets small fish invest in startups that need less than, say, $250k. Investments could have a lower limit of $5k for example, and most software startups might only need 10 or 20 investors at $5k a piece.

Diaspora received a great deal of funding on Kickstart, but Kickstart specifically says it is not meant for investing. I think there might be a niche between well-connected, wealthy angels and simply asking family and friends for seed money. This would also alleviate the choice between risking your relationships with all your friends and family just to generate seed money. If you have a business idea, there should be an option to raise money in an open market at any scale.

I've also had problems with paypal, due to the fact I needed three accounts for three separate businesses/ partnerships. Even though I formed separate LLCs, Paypal apparently can only link even business accounts to a specific person, so when I had two accounts based off my name it started causing no end of problems. The account would get locked, I'd call and have it fixed, then it would lock up again. They were friendly, but it was definitely a pain.

If someone could take this concept a step further and let you create a self contained process that users could download and run just by clicking (like tasks in photoshop), I could see some uses:

- Some tech support situations where you have to have a user do x amount of steps on their computer that are the same for all users. Sort of like an automated Geek Squad.

- Sell a prepackaged GTD style organization system that creates all the folders for you in the right places, downloads files (pre-made budget spreadsheet for example) into them, etc. (trivial, but it's a pain point for people)

- Make a bunch of different productivity apps that mimic the steps a professional programmer/ photographer/ marketer etc does when they first setup a new computer (bookmarks, preference settings, etc.)

I think this is a great idea, but agreeing with other commenters minimum wage is likely far too low. My brother and I are fairly new to ruby on rails, and we often get stuck and left at the mercy of support forums and friendly programmers. Many times we'd gladly pay for an "office hours" type service, especially since one hangup can cost us a day or two... one model might be to try letting users post questions along with an ideal price range and let programmers bid to answer it, like a stack overflow for business users.

Maybe it's just me, but I could see many uses for a tablet beyond just reading in the bathroom or in bed. For any job where you're standing/ walking and taking notes (real estate agents, etc), it's awkward to try to open a macbook, and the iphone is too tedious.

Pair a tablet with a solid voice to text app, and it'd be pretty useful for lawyers, reporters, etc.