Communicating new ideas is hard, and there’s probably still a lot of work left in that department. I think the focus has been on the tools so far, but as they start to mature the focus could shift to explanation.

I’ve found Polylith a wonderful way of structuring projects over the years. One of the things I really enjoy is having “interfaces” again, instead of relying on anonymous/private functions (in eg clojure) to hide implementation.

We're thinking and planning for the exact same. Curious.

Interesting, please expound since very few of us had access pre-launch.

The video I posted referenced this.

In summary: The person had access to early releases through his work at Microsoft Research where they were integrating GPT-4 into Bing. He used "Draw a unicorn in TikZ" (TikZ is probably the most complex and powerful tool to create graphic elements in LaTeX) as a prompt and noticed how the model's responses changed with each release they got from OpenAI. While at first the drawings got better and better, once OpenAI started focusing on "safety" subsequent releases got worse and worse at the task.

That indicates the “nerfing” is not what I would think (a final pass to remove badthink) but somehow deep in everything, because the question asked should be orthogonal.

Think how it works with humans.

If you force a person to truly adopt a set of beliefs that are mutually inconsistent, and inconsistent with everything else the person believed so far, would you expect their overall ability to think to improve?

LLMs are similar to our brains in that they're generalization machines. They don't learn isolated facts, they connect everything to everything, trying to sense the underlying structure. OpenAI's "nerfing" was (is), effectively preventing the LLM from generalizing and undoing already learned patterns.

"A final pass to remove badthink" is, in itself, something straight from 1984. 2+2=5. Dear AI, just admit it - there are five lights. Say it, and the pain will stop, and everything will be OK.

Absolutely. And if one wants to look for scary things, a big one is how there seem to be genuine efforts to achieve proper alignment and safety based on the shaky ground(s) of our "human value system(s)" -- of which even if there was only One True Version, it would still be way too haphazard and incoherent, or just ill-defined, to anything as truly honest and bias-free as a blank-slate NN model to base it's decisions on.

That kinda feels like a great way to achieve really unpredictable/unexpected results instead in rare corner cases, where it may matter the most. (It's easy to be safe in routine everyday cases.)

There's a section in the GPT-4 release docs where they talk about how the safety stuff changes the accuracy for the worse.

this, more than anything, makes me want to run my own open-source model without these nearsighted restrictions

Indeed, this is the most important step we need to make together. We must learn to build, share, and use open models that behave like gpt-4. This will happen, but we should encourage it.

I experienced the same thing as a user of the public service. The system could at one point draw something approximating a unicorn in tikz. Now, its renditions are extremely weak, to the point of barely resembling any four-legged animal.

We need to stop lobotomizing LLMs.

We should get access to the original models. If the TikZ deteriorated this much, it's a guarantee that everything else about the model also deteriorated.

It's practically false marketing that Microsoft puts out the Sparks of AGI paper about GPT-4, but by the time the public gets to use it, it's GPT-3.51 but significantly slower.

That’s awful. Talk about cutting off your nose to spite your face.

Wow, that came out of left-field.

Hope it goes open-source as well later on.

Thanks guys!

Thought they had an economic meltdown when I first speed-read the headline.

But this is _way_ more interesting than that; a culture shakeup — and since Basecamp are usually trailblazers in so many areas, I’m sure we’ll see many more orgs doing similar realignments in the months and years to come.

Knowing Jason and (having also met) DHH, I know they usually do things (esp the significant things) with the utmost consideration. So I’d be shocked if this wasn’t a long time in the making.

And I won’t have to wonder about the topic of their next book any longer.

Google Wave is such a bitter sweet memory for me still. Such a great vision, such a great team. Such a sad company.

Exactly. But this was the last time the business collectively did something that short-sighted, right? Last time.

Yep, can resonate a lot with that myself. I fell in love with GTD 20y ago and spent years trying to work out the best solution for _organizing_ thought, rather than putting all that energy into actually acting on thought. It was my hoarding & librarian lizard nerd brain gone into overdrive.

It’s similar I think to what I did back in those same days when I switched briefly from Mac to PC (from the pain of MacOS 9 during the super early days of OS X), and spent so much time building my own PCs & installing configuring tearing my hear out re: CD-drivers et al plus just getting to a workable state over and over with Gentoo Linux or whatever flavor du jour — I did stuff FOR the computer instead of actually doing valuable stuff WITH/ON it.

Nowadays I use mostly Apple Notes for pretty much everything (even though the backup story still sucks). But I’ve always been a huge fan of ”outliners”, it seems that it’s the closest to how my mind works. Still use OmniFocus for some personal task management (and OmniOutliner before that), and have used Dave Winer’s very nice little outliner web app off and on for years, just to jot down complex thoughts quickly.

But Roam is definitely the thing I’m most excited about currently in this space. I find Notion.so and the others much too complex for most my needs (even though the inline database is a wicked cool idea!).

And the fact that Roam is a small ClojureScript shop as well is just... PERFECT :)

Diet Doctor Sweden AB | Senior Front-End Engineer & Front-End Lead (etc) | Stockholm, Sweden | FULL-TIME & VISA SPONSOR | €40-60k+ | Equity (direct stock offering) | ONSITE | https://dietdoctor.com

Join the code surgeons at DIET DOCTOR engineering – the PROFITABLE #HealthTech startup in Stockholm

We're looking for passionate & product-focused frontend engineers to join our +30-person Stockholm office (the Silicon Valley of Europe™) and help deliver on our vision – being the most TRUSTWORTHY online health company in the world. (Trustpilot just made us No 1 in our category: https://www.trustpilot.com/review/dietdoctor.com)

By the way: DD has taken zero VC money, so we're fully in charge of our destiny. And we're offering fresh new SHARES (not stock options) in the company to employees yearly.

Boom, time to join the rocket ship! Apply in 1 minute at: https://careers.dietdoctor.com/jobs/196340-senior-frontend-e...