Greg Leppert

Invest in the millennium. Practice resurrection. Executive Director @instdin, Chief Technologist @BKCHarvard

[Greg Leppert @leppert]

When I feel intimidated learning new things, especially systems, I often think about “Fog of War” mapping in RTS games. A FoW map is obscured in darkness, only to be revealed as you explore, tile by tile, until the entire domain is visible.

[Greg Leppert @leppert]

@ericstromberg This year I'm hoping to take home a giant inflatable basketball.

[Greg Leppert @leppert]

@timhwang I’ve been saying this and that for years.

[Greg Leppert @leppert]

You can read the full brief here: https://t.co/RikZ1kDf1E

[Greg Leppert @leppert]

On Friday, I filed an amicus in the OpenAI copyright cases. These cases are immensely important to the future of creativity and AI in the U.S., and of libraries along with them. I shared with the court my belief that AI systems have a key role to play in the next era of libraries’ stewardship of information, and that a ruling against fair use for AI training could severely limit their ability to foster access to knowledge.

Information generation is accelerating, stewardship considerations are compounding, and libraries rarely receive the support they need. At current rates, it will take 100s of years—if not a millennium—for libraries to extend digital access to the whole of their collections. At the same time, we are finding new meaning and value in information, putting increasing pressure on our institutions to catalog, preserve, and provide access to it. What we are asking of them as stewards of our common knowledge will require new tools to augment their already immense efforts.

AI systems allow libraries to process collections in new, accelerated, and increasingly deep ways—from OCR to classification and description, through to search and discovery. This work requires general-purpose models that can work across a wide range of materials, including those under copyright, and the ability for libraries to tune these models to their needs. This is the essence of our work at the Institutional Data Initiative, but it is only possible where the law permits it.

I believe we are on a path to what I define as “ubiquitous” knowledge access: the ubiquitous availability of all knowledge and the ability to near-instantaneously traverse all possible connections between all that is known. As with any mythical story of knowledge, this is not a vision guaranteed to be utopian, but I do believe it to be an empowered one primed for thriving.

The library community is fortunate enough to be a cord made of many strands. My vision of knowledge access, and the role AI plays in it, is not shared by all of them, and we are better for it. Diversity among libraries has always been a strength. The same is true for the courts and the wealth of amici filed in this case, many of which represent views contrary to mine.

Authors and other creatives are under immense pressure, and it is equally important that they receive full consideration from the courts so that their works receive all protections to which they are entitled. The process of doing so is likely to highlight where we need to consider other forms of legislation, remuneration, or incentives to ensure they, and society, continue to benefit from their work. We would do well to take note.

There are no sides here—only our collective society. I hope libraries are able to thrive in it, authors and other creatives along with them, and that AI contributes to finding a path that enables that thriving for us all. I am hopeful all of these things can be true.

[Greg Leppert @leppert]

This is a great idea from a great builder.

[Allan Yu @allanyu_]

I want to introduce 𐌉𐌃𐌋𐌄 𐌋𐌉𐌍𐌊𐌔

One thing we’ve learned from our users is how quickly a simple task becomes a second job for the organizer.

Collecting $20 for your pickup league turns into a spreadsheet and chasing people. Selling something means a dozen DMs. Picking a date means scrolling through everyone’s availability. People don’t want 20 apps for 20 small things. They want a shareable link that does the job.

You’d think AI would help, but a personal assistant is different from wrangling people. And vibe-coding an app is a lot for these tasks.

With Idle links, just say what you need and get a working link in seconds, ready to drop straight into your chats.

People can pay, buy, sign up, or vote. No app to download, no agents required.

Idle tracks who’s done what and can text you updates so you don’t have to piece it all together. Keep easy jobs easy, available today!

[Greg Leppert @leppert]

Please listen attentively while I describe my (dream / vibe-coded app) from last night.

[Greg Leppert @leppert]

Also true for certain data releases, etc.

[Greg Leppert @leppert]

We are entering an era where software releases are simply cached compute.

[Pirat_Nation 🔴 @Pirat_Nation]

AI is now being used to recreate Adobe’s software, a developer has released seven free, open-source apps as clean-room alternatives to Adobe’s main tools:

- PhotoCraft → Photoshop
- VectorCraft → Illustrator
- FilmCraft → Premiere Pro
- LightCraft → Lightroom
- EffectCraft → After Effects
- DesignCraft → InDesign
- PrintCraft → Acrobat Pro

Built in Rust with help from Claude Opus 5.5, they run on Windows, Linux, and macOS under MIT/Apache licenses without using Adobe’s source code.

[Jonathan Zittrain @zittrain]

I wrote about this as “intellectual debt”

The Hidden Costs of Automated Thinking https://t.co/l45ugWONPC

[Nathan Lambert @natolambert]

Today we're unveiling Trillium Labs @trillium_labs, a new non-profit to foster the open science of frontier AI. We're building open post-training recipes and will expand into open infra to study RSI, reward-hacking, multi-agent systems, and whatever comes next.

We're built around the theory of change that you need more eyes to solve hard technical problems. We have faith in the scientific methods and communities that humanity has built, and worry that AI is becoming too closed to utilize them.

Trilliums are wildflowers that bloom briefly in the spring, before the forest canopies fill out. Though they are small, they lay the foundation for the cycles of growth and nourishment through the rest of the year. At Trillium Labs, the recipes will be the slow nutrients for the seasons and the model releases will be the blooms. Building an institution dedicated to this is needed because, much as nature’s trilliums are slow to expand and grow, the open-ecosystem needs time and dedicated resources to catch up.

I co-founded with with a long-time friend and collaborator Tom Zick (@thesezickbeats). We're hiring (full time + student collabs/interns), we're fundraising, and we're looking for compute. Please get in touch if you're interested in helping out. Offices based in the Bay Area and Cambridge MA, remote okay.

I’m in the Bay Area until for The Curve and COLM to connect with people who are interested. We’re thankful to have initial support from Halcyon Futures and Schmidt Sciences with more funding en route to enable our ambitions of scaling. Our advisors @Thom_Wolf, @HannaHajishirzi, @gneubig and @ctnzr have been instrumental to building the ecosystem that exists today, and I’m stoked to get to keep working with them.

[Greg Leppert @leppert]

SHEIN puts a CAPTCHA in front of marketing-list unsubscribe. Wild.

[Greg Leppert @leppert]

AI seems to have ushered in a golden age of analogies.

[Greg Leppert @leppert]

@maiab https://t.co/jcB1dkBdWQ

[dylan @narrenhut]

Breaking: the average person has terrible taste and likes crap that sucks

[Greg Leppert @leppert]

The best way to prepare for all possible AI futures is to invest deeply in your human relationships.

[Greg Leppert @leppert]

@jstn AI is all about relationships.

[Greg Leppert @leppert]

This and the Elon/Altman AI generated videos are proof that committing to the bit is the lion’s share of projecting real presence.

[Nick Khami @skeptrune]

can't believe palmer and pg get along now

[Greg Leppert @leppert]

It’s still a dream of mine to get featured on Surfstation and K10K.

[Web Design Museum @WebDesignMuseum]

In 2001, we really thought this was what the future of web design would look like.

[Greg Leppert @leppert]

@jstn <proceeds to dump every one of my creds onto X>

[Greg Leppert @leppert]

Then again, maybe she was truly a dangerously sweet child of his.

[Greg Leppert @leppert]

Beginning to suspect that whole Guns N’ Roses tough guy thing was just a front.