I looked into these guys recently. I do think their heart is in the right place. Like it's obviously a small operation, but I think that's all the better. We don't need another Mozilla foundation. But being small their marketing is a bit sparse and engineer:y so I'll take the opportunity to shill a bit.
They do sort of seem to worry about a lot of the right stuff.
As programmers and engineers we can have an idea and then through some effort just make that thing exist. It's really is tantamount to wizardry.
Sadly it feels like there are more obstacles to this every day, both in terms of literal hurdles like locked down hardware and app stores, but also the (erroneous) notions that you need to host your software in the cloud or use FAANG-scale solutions like kubernetes, or just the Jenga towers of half-baked libraries that make up the technology stacks many developers are stuck with means constant attrition from library churn any time you get anything to work.
And also this weird mentality that's drilled into a lot of inexperienced developers that they need to lower their ambitions, and to stick to building simple things.
I think that sticks with people even as their experience and capacity to produce software grows.
I think it's gotten to the point where many people with the skills to build amazing software just don't understand that this is something they're able and permitted to do.
You get to build software for the sake of software existing and providing a benefit to you, to yourself. If you need something, you get to build it. Software can be used to change your world, the world, it can be an extension of your will.
Programmers obviously get that they can write software, but not that software can be very impactful, it can change lives, even reshape societies.
You don't need to limit yourself to todo apps, text editors and static site generators. You don't need a license or approval or permission to do build crazy ambitious stuff for the sake of crazy and ambitious stuff to exist. Not yet anyway.
> I looked into these guys recently. I do think their heart is in the right place. Like it's obviously a small operation, but I think that's all the better. We don't need another Mozilla foundation. But being small their marketing is a bit sparse and engineer:y so I'll take the opportunity to shill a bit.
Mozilla Foundation has large overhead on management (CEO getting more while people getting fired), and it has that structure of a non-profit within a for-profit of the same name/brand.
For organizations providing grants, I don't think Mozilla even does that? Unless you mean stuff which is brought under the umbrella of Mozilla. Even YC and Google provide grants, its just with a twist (not selfless with the kicker being private equity). For an organization which provides selfless grants, there's for example the Dutch NLnet Foundation [1]. Which also falls under a different jurisdiction which can be a pro or con.
I'm contrasting them to the Mozilla Foundation specifically in the sense that it's very administration heavy, and produces an insane amount of hot air and marketing fluff (that aligns relatively poorly with its actions).
FUTO seems diametrically opposite in that it's ostensibly run by a small number of tech people.
NLnet is cool too. They seem relatively similar to FUTO in terms of goals, but they've been around for decades and it shows. Dunno how to describe them other than that they're very streamlined. Still also very low-nonsense and the people you interact with as a grantee are technologists and not bureaucrats.
Louis Rossmann works there now, and he has put out a number of videos about them over the past few months. The first couple of minutes of this video gives a nice overview:
Once men turned their thinking over to machines in the hope that this would set them free. But that only permitted other men with machines to enslave them.
Frank Herbert, Dune
Humanoids aren't 'symbiotic' with Minds. They are thoroughly unnecessary but the Minds keep them around for amusement and company. They are pets. It's a critical problem with his universe, even though it is a lot of fun to just ignore it.
Well, ok, I said symbiotic because I dodn't want to go into the weeds and yes strictly speaking it's not the right term, be it it's also not as simple as you imply.
There are humanoids that are assumed to be (at least nearly) as smart as some Minds, be it they number in the low tens, one of the reasons that Minds do keep biological humanoids around is because they're sometimes surprised by their (perhaps irrational) but interesting point of view, there's also a point of pride amongst the Minds as to who can be more emphatic towards the humanoids, so it's not as simple as you say, be it that wasn't my point.
It is, hope looks like is not enough, then you start doing something, again someone comes in to buy and monetize it. It became foggy with the currents. Maybe with the space in equation and infinit possibilities things may get exciting again.
I'm less worried about software freedom these days. Desktop free software has been solid for ages now, and we are starting to have SaaS replacements with the rise of the fediverse.
I'm worried about hardware, though. My wife got a new Dell laptop about six months ago, and despite my 20 years of installing Linux distros on every piece of hardware I can, it still took me hours to jump through all the hoops to get Ubuntu on the machine.
Microsoft is locking stuff down, hard. I'd be willing to bet that when Windows 12 releases, it will require that the PC be a closed platform that does not allow users to install another OS. MS sees people just living with the closed by default mobile ecosystem and wants something similar for PCs.
A lot of (almost all) software has dependencies on a handful of organizations. As ideology creeps into everything, I think we're way closer to lots of commonly used software being rendered difficult ot use for wrongthinkers. In particular most of the common machine learning / AI stack, which is subject to the current push to "pause", regulate, whatever.
Software and hardware freedom are inseparable. If you can't install free software because proprietary software locks you out, it doesn't matter that KDE is amazing nowdays.
Hence why free software i.e. Libreboot, free firmware etc. is imo as important as ever.
I agree, but I worry we're leaving the age where an average Joe with no particular computer skill can install free software. My wife, not really a computer person, installed Ubuntu on her old laptop by herself just fine. She didn't even tell me, I just walked by and saw her using GNOME one day. But for this new laptop, she had to call in a ringer.
Libreboot et al are important, but meaningless in the grand scheme of things if you have to be a serious computer geek to be able to install them on your own hardware.
The battleground is hardware manufacturers yielding to Microsoft's requests. Lock down the bootloader, lock down the bios, lock down the firmware, before you know it, it's as hard to install Ubuntu on a Dell as it is to install LineageOS on a carrier locked Samsung Galaxy.
We need manufacturers that say no to Microsoft and keep making unlocked PC hardware.
Sorry but I simply must ask; what were the hoops you had to jump through? I installed Fedora on my moderately recent thinkpad x13 (Ryzen 5850U w/ NVME, USB 3, and all the new fixings) and it effectively "Just Werked", about as easily as installing Linux has been on all my machines since core 2 duo-ish.
We had to work around all the Secure Boot bullshit. It wasn't terrible once we figured out that was the problem, but when we expected it to just work, it turned into a lot of frustrated searching around on the internet to figure out why stuff wasn't working.
I've found Ventoy [0] really great for automation of all those Secure Boot hoops. I remember my installation screen asking me if I want to setup the Secure Boot bullshit, I just said "yes" and everything worked fine.
Plus, it can hold multiple distribution images (including Windows) and let you choose at runtime which one you want to install. You just paste the ISO file onto the filesystem (which is NTFS or exFAT or something, so there's no file size limit). And since that's an ordinary filesystem, you can also use the spare space to move any files between computers as if it was ordinary storage drive.
I think my point still stands, though. Users shouldn't have to work around secure boot. Secure boot is making Linux adoption more difficult for casual users for no good reason.
If anyone has a use case for secure boot outside of an enterprise setting I'd love to hear it.
My sister bought a couple of low-end HP laptops last year, the kind meant for college kids. The damned things wouldn't boot w/o connecting to your MS account!
This is one of the most amazing organizations doing stuff today -- I wish I lived in Austin to participate more. I've visited a few times and am always impressed by the people I meet there (both the "staff" and the fellowship recipients); had the fortune of showing up during the second fellowship demo day and saw some great people/projects, a lot like early YC demo days.
A quick browse on your site leaves a few questions. You're based in the US but I couldn't see anywhere that it said you were open to approaches from outside the US. There is no description of the size of the grants that I could find, and there's no email to contact you. That information could allow me to rule myself out very quickly which is why I'm asking.
I saw your chat option but for privacy reasons, or paranoia if you prefer I tend not to give out my details to third parties, such as my email which you require to use chat.
If they are there and I didn't see them, that's on me. Thank you.
Yes! It's not on the site yet but we recently gave microgrants to VLC, FFMpeg, the Blender Foundation, as well as other non-US projects. We also just had a fellow[1, applications now open] from Lithuania. (I work for FUTO)
That's interesting, some time ago I saw a video about a live captioning system in Louis Rossmans (youtube id r09Hm2zd2lY) where someone asked the same question and the answer was that they only accepted local applications.
Did I misunderstand his answer or is this something that changed?
Asked him and he said he was talking about having to travel here to do the fellowship. "If you are international, we will accept you as a fellow and help you come to texas, but the fellowship program will not allow you to be in the fellowship program if you want to stay in your home country throughout the program."
I was also wondering and think it is good-practice if that is mentioned by any grant-offering institution. Institution? That is not mentioned either. There is only an address. Is it a non-profit, for-profit, who is working there?
We're building a new web browser (with a new engine) from scratch.
I believe the industry is heading down a dark path with even big companies like Microsoft and Opera hiding behind Chromium instead of continuing to develop their own engines. IMO we need more independent implementations of the web platform to ensure that it remains open and healthy.
Re: too ambitious, it's a commonly held belief that building browsers is impossible without a huge budget and hundreds of staff. We're doing it anyway, it just takes time. :)
I would also want to know this prior to any interaction. There is still so much money in this space, and the risk of your innovation being harvested / co-opted / subverted along with it. At this point Mozilla seems to exist solely as an anti-trust insurance policy for Google. How are we to know if FUTO is any different?
FUTO makes sense in terms of it comes from "FUTurO", in the context of Future, but I see the organization changing its name if it is going to become a large player - just having "FU" as the first letters is a bit unuserfriendly (although I realize that they probably want to signal that they are saying FU to the other companies they are competing with), and I won't even mention that it is oddly similar to "FUPA", maybe the founder(s) wanted to use this as a joke of some sort? Similar to Elon combining his model letters into "S3XY".
Officially it's just 4 letters that sound nice together... but the name is up for interpretation :D
My personal favorite is "Fight to Upend the Tech Oligopoly" but there are many others.
I think their hearts are in the right place, but IMO most of the concerns about user freedom are obviated or transformed by the existence of these new talking machines.
(Just in the interest of disclosure, I'm a "Free" software partisan (I don't see the point of "Open" source.) I don't want to argue about it, I'm just being upfront about my own bias.)
It seems to me that we have three great UI paradigms: CLI, GUI, and now LLM. The first is typically seen by normals as inscrutable and esoteric, the domain of computer "wizards". The second is "computer" in the minds of most normals, this is how you get University courses on e.g. how to operate MS office software, etc. The third, speech, is of course learned automatically as children.
From now on the kind of programming we know (and love, some of us) is obsolete. No one wants to operate symbolic calculators via janky teletype machines. No one really wants to learn boring video games like Excel or Word. We want to tell the computer what to do and have it do that. Well, now we can.
But that means that we, the hackers and computer nerds, are now playing with model trains. (That's my metaphor: am I writing code people will actually use? If not, then I'm playing with my model trains. Nothing wrong with that! Fantastic hobby! But I'm not starting a shipping company with N gauge trains, yeah?)
It seems to me that now, in 2023, most needful software has already been written, just as most needful technology has already been invented, and the focus of an organization like FUTO should be on simple open hardware, an open library of algorithms and DBs, and a simple LLM UI, all on top of something like Glamorous Toolkit. Plug that into an open IoT, declare victory.
FWIW, I don't know why they downed you either. Just thought that your epitaph was premature. Despite recent developments, the symbolic calculator via teletype is still getting me to the point of "correct" faster than ChatGPT. If I owned a search engine company, or wrote clickbait listicles for a living, I'd be worried, but I don't.
Natural language is packed with ambiguities and unspoken/unwritten context, and is thus a terrible programming language. People will end up spilling paragraphs of prose into text models, to do poorly and unpredictably, what a couple lines of bash would have done perfectly. For "approximate", natural language is fine, and possibly even optimal, but for precision, for "correct", it is unsuitable. It is why every profession develops its own domain-specific vocabulary and re-definition of phrases: they need more precision. Don't get me wrong. Those paragraphs will most certainly be spilled, but this is not the demise of programming. This is the next degeneration of programming! The Netflixification of programming; the union of hacks and technology, generating infinite low-quality content for insatiable consumers who have lost all faculties of discrimination.
Perhaps the "most needful" software has already been written. But how much of it is any good? Last I checked, everyone's bug tracker is still overflowing. Some are easy fixes. Others will be decade-long slogs. I wish we were as close to done as you think, because I think with these new models, we're actually going to start slipping further away.
> What does that have to do with anything?
If the market size serves an estimate of work the world wants performed, and something as "done" as a bicycle still pulls-in $68B/yr, while software still earns trillions, then by the money-metric we have quite a ways to go before we are anywhere near finished.
It's possible. I gave up prognostication when I didn't predict Twitter. I recently checked Craigslist for jobs in my line and it was empty...
> the symbolic calculator via teletype is still getting me to the point of "correct" faster than ChatGPT.
I prefer it myself. But we're not the mass audience, eh? My entire career is built on the fact that most people won't even learn Python or JS (let alone bash.)
When we connect these models to empirical feedback devices they'll be able to tell us correct things.
Schmidhuber says his goal is to "build an automatic scientist and retire."
> Natural language is packed with ambiguities and unspoken/unwritten context, and is thus a terrible programming language. People will end up spilling paragraphs of prose into text models, to do poorly and unpredictably, what a couple lines of bash would have done perfectly.
But you have to learn bash first. Now we can offload all that overhead to (fast powerful) external processors.
> For "approximate", natural language is fine, and possibly even optimal, but for precision, for "correct", it is unsuitable.
Nyet. Read the links I gave you. Our task is selection from the choices the machine generates. That's how you build an intelligence amplifier. The machines are capable of arbitrary precision.
> consumers who have lost all faculties of discrimination.
That's the only faculty the machine can't replace. That's the purpose of humanity: to answer the question "What are people for?"
I don't know what's going to happen, but I am pretty sure of the overall envelope of possibility. We might become the slobs in Wall-E, I don't know.
- - - -
> If the market size serves an estimate of work the world wants performed, and something as "done" as a bicycle still pulls-in $68B/yr, while software still earns trillions, then by the money-metric we have quite a ways to go before we are anywhere near finished.
I'm still not sure I'm getting what you mean? What's "finished" mean here? People buy bikes, yes, but I don't see the connection to my thesis that talking computers will rapidly make obsolete things we programmers take for granted now: Programming Languages, Bug Trackers, Bugs, Frameworks, Libraries, Apps, etc. Just as the smartphone subsumes many previously-separate devices, these "Talking UI" systems will subsume most of what we think of as programming.
> Nyet. Read the links I gave you. Our task is selection from the choices the machine generates. That's how you build an intelligence amplifier. The machines are capable of arbitrary precision.
I'm already familiar with the Engelbart paper, but this is what people have already been doing with Google; with copy-paste from Stack Overflow. More often than not, all of the generated choices are either partially or wholly incorrect, and in need of either a cross-breeding or a euthanization. In this respect, LLMs are no different from the answer generators which preceded them. Only the priesthood of RTFM knows what to breed, what to gene-edit, and what to put down.
Excellent! I hope I didn't come off too patronizing?
It's hard to argue this point (that the machines will rapidly become empirical, aka scientists) because I don't know how it will happen, it's more than a gut feeling, but less than an equation or a working example. I'm certain it will, but less certain how swiftly.
In a sense, I'm trying to get out ahead of things to the extent that's even possible. FWIW, I would bet the inflection point comes within five years, the computers become automatic scientists, the options they generate will all be non-hallucinatory, and the only job left for us is "be human".
They do sort of seem to worry about a lot of the right stuff.
As programmers and engineers we can have an idea and then through some effort just make that thing exist. It's really is tantamount to wizardry.
Sadly it feels like there are more obstacles to this every day, both in terms of literal hurdles like locked down hardware and app stores, but also the (erroneous) notions that you need to host your software in the cloud or use FAANG-scale solutions like kubernetes, or just the Jenga towers of half-baked libraries that make up the technology stacks many developers are stuck with means constant attrition from library churn any time you get anything to work.
And also this weird mentality that's drilled into a lot of inexperienced developers that they need to lower their ambitions, and to stick to building simple things.
I think that sticks with people even as their experience and capacity to produce software grows.
I think it's gotten to the point where many people with the skills to build amazing software just don't understand that this is something they're able and permitted to do.
You get to build software for the sake of software existing and providing a benefit to you, to yourself. If you need something, you get to build it. Software can be used to change your world, the world, it can be an extension of your will.
Programmers obviously get that they can write software, but not that software can be very impactful, it can change lives, even reshape societies.
You don't need to limit yourself to todo apps, text editors and static site generators. You don't need a license or approval or permission to do build crazy ambitious stuff for the sake of crazy and ambitious stuff to exist. Not yet anyway.