Chisel and this project are different; chisel is a programming language to design hardware while this is a build tool for hardware design.
In software we have a nice compilation process that transforms code into machine code. However, to "compile" an ASIC you go from a hardware description language like verilog (which basically describes the design) and pass it through a complicated pipeline that composes a number of different tools. Right now, engineers use ad-hoc flows that use bash/TCL to glue together all the parts --- the project posted above is an attempt to cleanly specify + control the "compilation" process.
No, it it is not a competetitor to openroad. It's more like a build system, like how make sits on top of llvm. Openroad is one of the tools under the hood.
Not sure if I follow your JFT argument, but there's a large body of work on both (a) studying whether chasing ImageNet accuracy yields models that generalize well to out of distribution data [1, 2, 3] and (b) contextualizing progress on ImageNet (i.e., what does high accuracy on ImageNet really mean?) [4, 5, 6].
For (a), maybe surprisingly the answer is mostly yes! Better ImageNet accuracy generally corresponds to better out of distribution accuracy. For (b), it turns out that the ImageNet dataset is full of contradictions---many images have multiple ImageNet-relevant objects, and often are ambiguously or mis-labeled, etc---so it's hard to disentangle progress in identifying objects vs. models overfitting to the quirks of the benchmark.
> I've seen TikTok videos go viral being really critical of the US government including its foreign policy. The kind of video that would just be invisible or outright banned from Youtube or Facebook video.
Do you have any evidence of this happening? I've never heard of pro-US censorship before on any of these platforms (or any social media platform..)
I'm struggling to understand your reasoning. If the author had said "I only drink fair-trade coffee" would you also object if it happens to be the case that the fair-trade coffee is the cheapest type at his local supermarket?
We're discussing an educational institution with power and authority (MIT), which promotes drinking "free coffee", whilst simultaneously portraying it as somehow morally superior to "coffee one has to pay for" to consume.
Growing and selling coffee takes time, labour and effort - yet none of that is being reflected or accounted for when we choose to not pay for the coffee we consume.
Is this a sustainable approach? Does it promote "choosing the right tool for the job", or does it promote blind idealism ("free is better")? And why does a university, which takes exorbitant tuition fees, not prioritize the best software for the course (over the one that's merely free)?
You seem to fail to grasp the difference between "free as in beer" (gratis) and "free as in freedom" (libre).
Your coffee-response is all wrong: this is not about "free" coffee, about (not) having to pay for coffee, but about a moral stance on coffee: e.g. demanding the cafetaria only serves Fair Trade coffee, regardless of the price.
This is not about "having to pay for it" at all. The opposite really: running your own jitsi or BigBlueButton is probably more expensive than using the free tier of Teams, Zoom or Hangouts.
I run a large-ish jitsi instance: approx €50/month just for the VPS, my hours probably add another €2000/month to that.
> "You seem to fail to grasp the difference between "free as in beer" (gratis) and "free as in freedom" (libre)."
Why are we going back-and-forth about this again?
I previously asked whether there were concrete examples of software used during the course that were "free as in freedom" but NOT "free as in beer", and the response to my question was that no such examples were available.
It's the actions that matter here, and the bottom line is that they weren't intending to pay for anything to begin with. And by that, they were abdicating the "it's not about the money" argument, in my honest opinion.
> "I gave you an example by quoting how expensive it is to run a "free" jitsi service."
That's an entirely meaningless argument. Here's what you're doing:
> "I run a large-ish jitsi instance: approx €50/month just for the VPS, my hours probably add another €2000/month to that."
Here's what Sussman is doing:
> "I used a Jitsi Meet server that I installed on an obsolete and otherwise useless computer that was sitting idle in my laboratory, on its way to the electronics junk heap."
The two scenarios are not comparable.
I'm not making exaggerated claims here. The University can absolutely afford to do better, the students (who pay exorbitant tuition fees) deserve better, and any "libre software" idealism here is simply people trying to cut costs, jeopardizing the quality of education and the overall experience, while touting moralistic superiority...
Distance learning, for example, could've been a much more widespread and accepted thing, had it not been for instructors cobbling up together scrapyard-bound hardware to use as a chat server. Coming up with a proper solution takes investing (time, money, expertise) - which some people will evidently avoid at all costs...
"It made available licenses for various nonfree programs, but I objected to them on grounds of principle.".
Which is to say Sussman could effectively, for the purposes of this course, get _any_ license for free-as-in-beer.
So the actual decision would purely be on some other grounds. Since Sussman is a world renowned CS teacher, I choose to believe he made his choice based solely on whether it was most suitable to teaching CS.
(This is not an unreasonable belief: The concept of "Free Software" guarantees that the student is able to take the software apart to see what makes it tick. That is obviously a very valuable property when learning how things work!)
> (This is not an unreasonable belief: The concept of "Free Software" guarantees that the student is able to take the software apart to see what makes it tick. That is obviously a very valuable property when learning how things work!)
And something that resonates very well with the basic attitude and culture of academia/research.
The article you point to illustrates that that basic attitude and culture is under threat. I agree. Openness, being able to build on the works of others, and learning "how something ticks on the inside" are still basic to science though.
I, too, wish that what the article illustrates weren't happening, but are you really arguing that you can't take the idea from a published paper, understand it and build your own work on it? The FOSS philosophy is the closest equivalent for code.
> promotes drinking "free coffee", whilst simultaneously portraying it as somehow morally superior to "coffee one has to pay for" to consume.
The price does not enter into Sussman's argument. Maybe it could, given how students tend to be short on money and abhor paying for expensive textbooks, but it doesn't.
> Growing and selling coffee takes time, labour and effort - yet none of that is being reflected or accounted for when we choose to not pay for the coffee we consume.
In my coffee analogy, the promotion of the fair-trade alternative would be based on the fair-trade mechanism for ensuring the lower levels of the production chain receive a fairer share of the income. Whether or not the author pays less or more at the store is irrelevant to the argument.
> Is this a sustainable approach?
For this class? Almost surely!
For some people (e.g. me)? To a large extent (things aren't black and white). Apart from (the admittedly large chunk of) non-free Javascript run by the websites I visit, and some firmware, my computing world runs entirely on FOSS.
For absolutely everyone in every situation? Surely not. That's OK.
Really, the only point that matters a lot here is the first one.
> Does it promote "choosing the right tool for the job", or does it promote blind idealism ("free is better")?
It seems to me to also promote the fact that FOSS is far more compatible with academic culture and behavior. While indeed you may have to pay publishers for access to articles (luckily a practice that's on decline!), you certainly have complete freedoms to build on the work presented in those articles for your own research!
I'd go so far as to say that no closed tool can ever be "right for the job" in an academic research setting! (Although one sometimes does have to compromise when no adequate alternatives exist, especially when it comes to lab equipment – but in the CS world things are a lot better.)
> And why does a university, which takes exorbitant tuition fees, not prioritize the best software for the course (over the one that's merely free)?
I don't understand what MIT's tuition fees have to do with this.
> "In my coffee analogy, the promotion of the fair-trade alternative would be based on the fair-trade mechanism for ensuring the lower levels of the production chain receive a fairer share of the income. Whether or not the author pays less or more at the store is irrelevant to the argument."
But for there to be any form of income trickling down production chains, someone has to be willing to pay something for the services they consume. Is this controversial in any way?
> But for there to be any form of income trickling down production chains, someone has to be willing to pay something for the services they consume. Is this controversial in any way?
No, but it is also entirely orthogonal to the blog entry we're discussing. At no point does the monetary cost of something enter into what Sussman is talking about. You seem to be conflating free-as-in-beer with free-as-in-freedom (man, I feel nostalgic writing those phrases again – I don't mean this as an offense, but the difference seems to be so much more well understood among the tech savvy these days than it was 15 years ago).
This is where the coffee analogy becomes relevant. While indeed the cost of coffee affects a lot of things in the coffee value chain, the idea of fair trade (for all its flaws, let's not digress into those here) are not about the end product's cost.
Your question is missing the point, but to give an answer:
Sussman taught the first edX course, 6.002x. Agarwal took credit for it (since he shot the videos and was the face), but Sussman did the plurality of the work, followed by Terman, by Mitros, and then by Agarwal.
Open edX is free-as-in-beer but not free-as-in-price.
I'm a computer science PhD student, and I mostly maintain/create open source projects (https://github.com/lengstrom/) in the evenings for fun --- its particularly great you're working with friends!
I've never viewed it as a potential full time job. However, I recently got GitHub sponsors and make about 5$ a month!
Hi HN, author here! As students facing online instruction, we built ZoomerBackgrounds to solve firsthand the everyday problems of having subpar backgrounds and spoke with several friends working in industry about their background needs. The result is a product we really think you're really going to love. We had a lot of fun making this over the weekend and hope you all enjoy :)
For the tech stack we used react on the front end with a Node.js server + Firebase for persistent storage. This was my first time writing Javascript in about 5-6 years and it's really neat to see how the ecosystem has matured - there are even real classes now! We also used the Google Cloud Vision API for detecting inappropriate images.
On the database side, we found that Firebase Store works really well until it doesn't -- when you want to do anything that isn't basic reads/writes (e.g. SQL-like IN queries, query pagination, programmatic queries like sorting by the HN ranking algorithm ~ points/(time since post)^(1.5)) the workarounds can get complex and a little weird.
One question we had is around AdSense approval! The Google Adsense approval team rejected our website for not having enough meaningful content / not having "valuable inventory" - it would be great to hear some insight about how to show Google we are a real website.
I see that you have an option to not include a watermark on downloaded images, but having watermarks turned on by default (especially since these look like user uploaded images which I'm assuming you don't actually own) is quite off-putting.
I was going to say that the watermark is too obtrusive ... I don't have a problem with you getting a bit of attention in exchange for the time it takes to curate the collection but it's too big and in a bad spot.
We took a look at adversarial examples for linear classifiers (and in general, we looked at properties that adversarial training induces) here: https://arxiv.org/abs/1805.12152 For $\ell_\infty$ adversarial examples on linear classifiers we found that adversarial training forces a tradeoff between the $\ell_1$ norm of the weights (which is directly associated with adversarial accuracy) and accuracy.
It looks like this article works through something vaguely similar for $\ell_2$ adversarial examples. It would be interesting to compare the author's approach with explicit adversarial training.
Whoa, I made Hextris (https://github.com/hextris/hextris, one of the games removed from the store) a few years ago! Is there any precedent in OSS developers being held responsible for misuse of their code?
The licenses for most software and other practical works are designed to take away your freedom to share and change the works. By contrast, the GNU General Public License is intended to guarantee your freedom to share and change all versions of a program--to make sure it remains free software for all its users. We, the Free Software Foundation, use the GNU General Public License for most of our software; it applies also to any other work released this way by its authors. You can apply it to your programs, too.
When we speak of free software, we are referring to freedom, not price. Our General Public Licenses are designed to make sure that you have the freedom to distribute copies of free software (and charge for them if you wish), that you receive source code or can get it if you want it, that you can change the software or use pieces of it in new free programs, and that you know you can do these things.
Part of the problem with letting people have freedom is that they have the freedom to make decisions that impact communities in a negative way. But it's usually worth the tradeoff.
This is probably the most relevant line though:
For the developers' and authors' protection, the GPL clearly explains that there is no warranty for this free software.
i.e. you have nothing to worry about, but you also probably can't do anything to punish the misuse. After all, misuse is subjective.
In software we have a nice compilation process that transforms code into machine code. However, to "compile" an ASIC you go from a hardware description language like verilog (which basically describes the design) and pass it through a complicated pipeline that composes a number of different tools. Right now, engineers use ad-hoc flows that use bash/TCL to glue together all the parts --- the project posted above is an attempt to cleanly specify + control the "compilation" process.