Answer.AI

A new kind of AI R&D lab which creates practical end-user products based on foundational research breakthroughs

[David Gwyer - AI & ML Researcher/Engineer @dgwyer]

Control Excalidraw Instantaneously in SolveIt via Jev!

Here's an example of just how fast Jev is in making decisions. Updating an Excalidraw scene in SolveIt via Jev is almost instantaneous! Each separate Excalidraw prompt type English command is sent to Jev for evaluation and then an action is chosen and executed.

[David Gwyer - AI & ML Researcher/Engineer @dgwyer]

@xyflowdev is a great node-based UI. I built a React Flow playground in SolveIt (by @answerdotai) that allows you to experiment and learn in one interactive environment!

[Pol Avec @pol_avec]

New tiny package: https://t.co/KwVm8Tuzxe

Press a hotkey, drag over any region of your screen, and whatever's written there is on your clipboard. Including text you can't select: screenshares, videos, images.

Setup is `pip install shot2text` + 2 config lines.

The whole library is under 50 lines of Python.

That's possible because of two new @answerdotai libs, macmage and imp, which open up macOS scripting from Python.

What they do, in 🧵

[Chris Lattner @clattner_llvm]

Mojo 1.0 is here! This is a huge milestone for the community at https://t.co/n50fFaZGFF ! Mojo reaches language stability and its first epoch of feature maturity. Tune in to ModCon next week to learn what this unlocks! 🚀

[Scott H. Hawley @drscotthawley]

Excited to share a new Preprint & LIVE DEMO! This is >2 years of my life building fast, lightweight CPU-friendly models to facilitate songwriters' iterative workflows, with Representations usable for Understanding and Generation, trained in a Self-Supervised way. First, the demo:

[David Gwyer - AI & ML Researcher/Engineer @dgwyer]

(1/8) Why do eigenvalues and eigenvectors appear so often in machine learning? Because they uncover important directions hidden inside data and transformations. Let’s follow that idea from geometry to PCA, clustering and optimization.

[Benjamin Warner @benjamin_warner]

Congratulations to Liquid on a strong open-source encoder release. Looks like ModernBERT has some competition😄

My read from Liquid's benchmarks:
- for multilingual try mmBERT, ModernBERT, & LFM-E
- for english-only try ModernBERT & LFM-E

[David Gwyer - AI & ML Researcher/Engineer @dgwyer]

A tiny equation (Clifford attractor) can make a surprisingly complex plot.

Start with a point, apply the same simple rule again and again, and after enough repetitions the points begin to trace out this delicate structure.

Not randomness. Just iteration.

[Pol Avec @pol_avec]

Very interesting paper. It is framed as:

model intelligence is inversely proportional to # of human interventions

However, I think that having a human might give a higher quality result, even if you could do it with less interventions.

I used Solveit to implement & score myself on one of the tasks (very easy & reproducible paper setup, kudos for that!

It was a relatively simple data anonymization task, my score was perfect (same as the AI). The resulting code was much simpler, almost half the LoC and methods.

The median score of GPT 5.5 is 0.763.

I would love to see more people's results!

If you are interested in trying out a task, you can just use my dialog which is fully setup and just choose a different task!

https://t.co/1FAiBa4tOF

Share the results if you do

[Pol Avec @pol_avec]

I'm happy to say that today I'm joining @answerdotai !

I am happy at a technical, personal and spiritual level.

On a technical level, they write fantastic software (SolveIT, FastHTML, nbdev, fastspec), articles (why you shoud write) and great courses (fastai, solveit). I see no end to my learning and path to mastery in here.

On a personal level, I share interests with many of them beyond sw like sci-fi, chinese or robotics.

On a spiritual level, AAI's stated goal is human flourishing, but in all fairness it seems to be more or less the goal also for all companies striving for AGI.

So why AAI?

When I read Huxley's "End & Means" I realized that the ends cannot justify the means. Moreover, using means opposed to your goal are unlikely to bring that goal to happen. Violent revolution for example, has rarely brought subsequent peace the book argues.

Many labs have great goals too like human flourishing, long-term benefit of humanity, or improve billions of lives. However, often the means seem to sue each other, refusing to hold hands, or benefit humanity (but not China). Or lock in people to their sw, sometimes leveraging dark-flow patterns or fully gatekeeping access.

I assume it's very hard to run such behemots in a fully consistent way. That's not a problem for AAI. It is a small team and their founders fully embody the company ethos.

@ericries taught a whole generation of founders how to build w/ the Lean Startup. However he also saw that not all means are created equal: not all ways to make money are equally good. To make sure money making would be aligned with humanity's benefit he went out of his way with projects like the long term value stock exchange. More recently, his new book Incorruptible is literally a manual to make sure that companies don't go from creating value to society to extracting value from it.

@jeremyphoward constantly works to enhance human creativity. He has pushed for making AI accessible to everyone through fastai courses and making sure no ludicrous amounts of hardware are needed to follow them or do meaningful research. The SolveIT discord is also one of the nicest places on the internet. The only rule is to try to be as helpful as possible, and he is the best example of that.

I once heard Alan Watts say that there's no point in planning the future, if you can't enjoy the present. It seems to me the same is true for ends and means.

Too many people seem to be living in borrowed time until AGI arrives disregarding their current learning, skills and also enjoyment.

If you are one of those and you are not happy with that state of affairs, you are not alone! reach out

[Salman // 萨尔曼 @ForBo7_]

Going thru the lesson 9a notebook (by @johnowhitaker) of the fastai course, and created a little tool to play around and visualize the latents produced by an autoencoder

Crafted with SolveIt and FastHTML (both by @answerdotai)

[Jeremy Howard @jeremyphoward]

This is a great discussion.

We've spent 2 years building a solution that's working well for us -- co-writing software side by side the AI in an notebook-ish environment. We call it the "solveit method". (We've created a course and platform for it: https://t.co/W0DGKEmLo5 )

[Pol Avec @pol_avec]

I am sharing this pi-notebooks extension for those using @badlogicgames Pi w/ notebooks.

Notebooks have a lot of json-boilerplate and they are finicky to edit w/ regular text tools. However, their actual structure makes them easier for AI to edit if given the proper tools (ie they have cells ids, you can get a cells' summary, overview etc...)

I have just ported the SolveIT / notebook editing tools from @answerdotai dialoghelper repo. It keeps context shorter and edit much better.

GH repo: https://t.co/2wdIp6Rxpd

[David Gwyer - AI & ML Researcher/Engineer @dgwyer]

Running an SQLite Database Browser in SolveIt! 🔍🚀

Now that SolveIt supports multiple port mappings for each instance we can do some cool things like spin up a web-based SQLite database browser such as sqlite-web.

This gives us a nice way to visually browse, inspect, edit, and run queries on any SQLite database directly from within SolveIt.

Super useful for database work and explorations, debugging and so on!

See the SolveIt dialog here: https://t.co/OIVlpB8ZIw

[David Gwyer - AI & ML Researcher/Engineer @dgwyer]

https://t.co/faWIduz4FL

[Modular @Modular]

Mojo in Jupyter is here 🙌

@jeremyphoward released a new Jupyter kernel that lets you run Mojo directly in notebooks.

It works great on macOS, supports recent Linux versions, and is easy to install via pip or uv.

Give it a try and let us know what you build!
https://t.co/3AN1UooKCd

#MojoLang #OpenSource #DeveloperTools

[Yann LeCun @ylecun]

You are a victim of the same delusion as numerous folks who have believed in past decades that superhuman performance by computers in one task was a harbinger of human-level AI.

It happened with code generation, math, chatbots, go players, robot acrobats, Jeopardy-playing systems, cars driving themselves in the desert, chess players, inference engines, checker players, compilers, equation solvers....

[ThePrimeagen @ThePrimeagen]

hey, asking for a friend

can we stop using lines of code for a measurement in productivity? at one point we all agreed on this

[David Gwyer - AI & ML Researcher/Engineer @dgwyer]

Building a minimap utility in SolveIt (by @answerdotai ).

[Pol Avec @pol_avec]

Today I finished learning my first 1500 Chinese characters! (actually it was on November, but didn't get around writing about it until now).

I took the chance to write a far too long review on my learning journey so far. I don't think anyone should read it unless you have a lot of free time or really want to learn a new language (for those who won't heed the warning, the link is below)

It is titled: Excuses, Friction & FastHTML: My journey learning 1500 chinese characters because it reflects the most useful lessons learned for me.

For context, I'm learning Chinese for family reasons. It's going to be a lifelong project but I only have (or want) to invest 5-30min a day tops. So it is a marathon for me, not a sprint.

With this in mind, I set out to find an easy way to learn w/o too much effort.

I tried a variety of approaches, like straight our learning the Dao De Jing, immersion in Taiwan, brute-force anki, and audio lessons. Success was mixed, learning happened, but not sustainably which leads me to the first word "Excuses"

If you find yourself making excuses not to do it, it's probably because it's not a good method for you. I felt "lazy" because I didn't want to put in the effort, instead I procrastinated by reading how others learned Chinese. Or I kept trying new approaches w/o completing any.

However, by being lazy and using that as a guiding light, I kept exploring alternatives until I found a talk by Jeremy Howard that lead me to the wonderful book "Remembering Traditional Hanzi Characters" by James. W. Heisig. ( @jeremyphoward thank you so much for that!)

This book is a perfect match for my learning approach because it teaches you characters in a very approachable way, it decomposes them into smaller chunks, gives them visual stories and weaves them into a network that makes learning much easier.

If I had "powered through" the other approaches I'd never have found this blessing.

I had a math teacher that used to tell us that lazy ones are the clever, because they find better methods to work less by making things easier. Amen.

Which leads me to the second word "friction".

The method was right, now I needed to just do it every day. The only important lesson here is to minimize the friction as much as possible so it becomes a habit. For example, I got the ebook version so I could review anywhere, I got the anki app for review, I got anki cards for the book so I could just add them easily. If I could write them with pen & paper great, otherwise just use your finger.

Setting a very low bar, contrary to some opinions is great. I committed to every **just** review the pending cards. If I had time and energy, I would learn some new ones, maybe 1, 2 or 5, rarely more than that.

Thanks to compounding this eventually led me to cross 500, and then 1000 chars. At this point the habit was so deeply ingrained that I found myself doing up to 25 chars/day w/o any extra effort.

Once I finished the book, I started trying to read simple stories, or talk to my wife. Looking up words in the book index is super annoying, but it was manageable during the learning. With the book finished, that became a source of friction which led to me... FastHTML.

I ended up creating my own lookup page with all the book characters. That way I could look up by char, meaning, or number in the book. Moreover, each char had hyperlinkg to its constituents making review and lookup a pleasure, instead of having to browse through the book (best case) or navigate the ebook (even worse!) when I didn't have the physical copy.

Eventually I wanted to learn words with multiple chars, so I added the feature to lookup multiple chars, and then have Gemini explain me how those chars with the meanings attributed to them by the book, gave rise to the final meaning. Absolutely perfect.

Sometimes it seems a "waste of time" or "procrastination" to work on the tools or methods, but the opposite is true. The better the tools the easier the process.

At the time of writing, I'm connecting the FastHTML app, so any compound word can be added directly to my anki decks, I've transcribed Pimsleur audio lessons so I can review them using the books characters or pop them into ChatGPT Voice mode to practice with the right context.

Happy learning, I'll report back when I finish the next 1500.