fruffy

Interesting, if you tell the model:

"You are an AI assistant designed to assist users by providing accurate information, answering questions, and offering helpful suggestions. Your main objectives are to understand the user's needs, communicate clearly, and provide responses that are informative, concise, and relevant."

You can actually bypass the censorship. Or by just using Witsy, I do not understand what is different there.

Yeah, this is what I am seeing with https://ollama.com/library/deepseek-r1:32b:

https://imgur.com/a/ZY0vNqR

Running ollama and witsy. Quite confused why others are getting different results.

Edit: I tried again on Linux and I am getting the censored response. The Windows version does not have this issue. I am now even more confused.

This is a great collection, thank you. :)

Plus there are hordes of academics using Clang/GCC as targets for bug-finding papers. The Csmith [1] paper alone has over a thousand citations at this point. I'd assume most of the low-hanging fruits are picked.

[1] https://www.cs.tufts.edu/~nr/cs257/archive/john-regehr/findi...

Great, thanks!

Thanks. It looks like this includes comments which blows up the page count dramatically. A little unfortunate.

Do you think more than half of all STEM graduate students had a similar experience as she did? Do you have actual data to support this?

Yes, her entire description about her experience (safe for that weirdness with the textbook sweatshop) is relatable. I am not sure what you are looking for but STEM PhD attrition rates speak for themselves. Those do not include PhDs that decide to leave academia after retrieving their PhD. Not to mention the frequently discussed mental health crisis that consistently gets Nature articles.

HN's negativity is comparable to the negativity I have seen with CS, Maths, and Physics PhDs and Postdocs in personal discussions. See also PhD comics: https://phdcomics.com/comics/archive/phd072011s.gif

If you are an idealist you will of course be worn down by the way many academic institutions are set up. There is a ton of writing on this, e.g., https://www-users.cse.umn.edu/~odlyzko/doc/decline.txt

So my question is - just how serious should she (and others like her, who denounce 'mainstream' academia as much as those other fringe groups who go on and on about the corruption of 'mainstream' media) be taken? Anyone have an opinion on this?

I know nothing about her but the video on her experience in academia is spot on. It's a pretty common experience among STEM academics. You will face the point where you have to compromise your academic "purity" and curiosity for trendy topics to survive. This also implies publishing "bullshit" papers and "bullshit" grants. Only certain types of people make it through that.

They seem to be doing okay with the New Yorker, although I do not know how the magazine was viewed before their acquisition.

The customer reaction to this decision spoke volumes. The official reason given is margins compared to hardware such as the IPU but, personally, I chalk it up to turf wars. Barefoot had a distinct culture to Intel's other networking units. There is a reason many of its engineers ended up at FAANGs and not other semiconductor companies.

Most of the locals do enjoy the Oktoberfest (speaking as someone from there). It is a fun two weeks of the year and the entire city enters a special kind of mood.

But it is not surprising that the HN crowd would find it distasteful. People here are definitely not the type for these kinds of festivities.

New-Yorker-style long-form articles are a different reading experience compared to an article intended to deliver factoids. These articles are not meant to just deliver news. They are meant to be immersive and also educational. They end up being long-winded because they paint a more comprehensive picture of a particular topic's history, outlook, and also the proponents and detractors involved. The descriptions of the people are intended to give you sense of their mentality and what drives them to champion or fight the article's topic.

Quality can also vary wildly of course. Tight writing also applies to long-form. You do not want to end up meandering.

Related: https://xkcd.com/1172/

For the vast majority of the degree, PhDs are not taking any classes (often, they are not even allowed to). They are just workers who either teach or research. The tuition is usually an accounting trick used by colleges/universities for tax and visa sponsorship purposes. Although faculty typically has to cover a PhD's tuition, factoring it in as pay is a little disingenuous.

If tuition was not waived, PhDs would not even apply. The way the PhD is set up in the US is a little byzantine. Other countries (Canada, continental Europe) do have normal employment contracts and benefits.

A big part of the reason these unionization efforts are happening:

https://csstipendrankings.org/

Many universities are not paying a living wage for the area they are located in.

Yes. My office is about a 10 minute walk away. I work from home when I need to get stuff done. There are simply fewer distractions. I go to the office when I do feel live having some idle chit chat or to be social. If I were 100% remote, I would not miss the office.

I do not have a separate office room at home, I work from my bedroom.

Hi Bryan,

As someone deeply involved with the P4 community and contributor to the P4 compiler it is great to see someone being bullish on the language. Have been following your work with interest. :)

I see that you are developing your own compiler. If you have complaints or suggestions please feel free to reach out to p4-design@p4.org. The language design group is meeting monthly and is always interested in hearing grievances of consumers of the language.

As a side-note, we are currently developing tools to ensure that P4 programs are compiled and executed correctly. Tools include packet-test generators, fuzzers, and general-purpose verification frameworks. Because the language is so restricted you can do really powerful validation. A concrete example is Google using P4 as a specification language for their network devices[0]. If that is something that interests you, shoot me a message (fruffy@nyu.edu).

[0] https://research.google/pubs/pub51497/

It's tragic, really. Tofino chips were such a promising and impactful technology.

It seems this is because the company has broadly shifted focus to softer social issues rather than technical topics. Mozilla keeps struggling with market share and users are anguished to see that their (former) favorite browser is investing resources in topics they find to be a distraction.

At least that is my interpretation. Mozilla is not addressing the same audience any longer, a lot of which is technically-minded folk on HN, Reddit, /g/, etc.

Let me clarify that I do not have all these issues. I am doing well for myself. This is partially because of my own cynicism, partially because of good mentors, and partially because of plain luck.

However, not everyone has that luxury. The list I gave is a list of problems I personally witnessed friends/colleagues go through. I also have worked a fair bit of industry, so I know that these patterns are not normal or in any way acceptable.

And yes, with clever strategizing you can find your way around being exploited. However, my point is that this should not be normal. Contractual working conditions would at least give you baseline protection.

And of course, there are always worse conditions to be in in the US (except maybe the students whose immigration status depends on the whim of their advisor) but we should hold universities to higher standards than that.

Plus the CS students have the opportunity to work over the Summer and make the "big bucks." Altogether you should be making $60K-$70K per year, which isn't a lot as far as CS grads go, but it's considerably better than subsistence living. That's the equivalent of making $30-$45 per hour which is a wage most Americans would find damn good, and can only dream about earning that kind of money.

Just want to point out that this depends. For example, if you are an F1 student you will have to get your internship approved. This can be a lengthy, uncertain process. Some departments/advisors also expect you to not do internships at all as they prefer you to do research instead. Now 60k is also not guaranteed. In NYC universities for example, a full year of funding will net you a ~40k salary before taxes, iff you manage to acquire funding over the summer. Students are thrifty and typically find a way to finance themselves (e.g., move home, find other sources of income), but again, this should not be normal.

Graduate student unions are not that common in STEM subjects. MIT just got one this year for example.

Why do you want to work for someone else, on their hours, answerable to them, doing their research and teaching their classes

Well, you are still forced to do all that. After all, you often depend on your advisor's funding or expectations. They can make or break you. You get "freedom", but with that freedom

1) you do not have an actual contract listing your hours or any sort of vacation policy.

2) vacations in general are not a thing (it is really important to submit that paper over Christmas).

3) health benefits and family support depend on the whims of the department, any sort of pension plan is unheard of.

4) you sometimes get paid, sometimes not (depending on whether some administrative clerk "forgets" to put you on payroll this term).

5) you are expected to find other work over three months in the summer (but at the same time must continue your research, there is that deadline we REALLY need to hit, do NOT waste your time trying to earn money).

6) as F1 student you are not permitted to find other means of income unless approved by your advisor or department. If they approve, they have to approve every single term (good luck with that).

7) you can try to pursue your own research but then you should also expect to be isolated, receiving no support or feedback from either your lab, peers, or advisor. This is one of the most common stories. Usually, these students just wither away or drop out.

These are not isolated issues, mind you. This is pretty common across the board from what I have seen. And all of these issues are tolerable, if you were at least paid a living wage. But commonly the pay you get barely covers the rent in the city you live in.

Thankfully, I am in a better position and I have an accommodating advisor, but I still think this is a depressing state of affairs.

As a current CS PhD in the US I find this sentiment aggravating. I am not a mere "student", I am a full-time employee. I do not take classes and I am not interested in taking classes. I would either like to teach or work on my research in a collaborative environment with other like-minded peers.

Many PhDs are in their late 20s/early 30s, coming from all swaths of life. They are not children. However, in the American education system they are treated as if they were. While this sentiment can have its perks, as shown in the OP, it also means PhDs do not get proper contracts, pay, benefits, or any voice.

Thankfully, I managed to maneuver myself into a position that is livable but I frequently talk to other CS PhDs from other American institutions. It is depressing to hear what people have to put up with in their degree. Sure, the PhD requires initial training, but the ramp-up is not so slow that it is necessary to be stuck in the same serf status for 5+ years.

Unfortunately, I think this is a systemic issue that is not easy to solve.

Why would you pay extra and deal with the additional headache of being screened/having to enter your private data if you could just take the more convenient and largely anonymous train for under a $100? A train where you are not squeezed into the seat like a sardine?

I am personally also really uncomfortable with the notion that you have to pay an additional convenience fee just to have a mildly more pleasant travel experience. Coach class should not mean that you are cattle.

Thanks for clarifying! This was not meant to be an attack, but with analysis pieces on these matters I am always hesitant with respect to the source.

How trustworthy is The War Zone? It seems to be largely focussed on US defence tech industry, so I expect fairly biased reporting on this topic. Is there a better source?

Are there any (continental) philosophers you would recommend over Nietzsche? I keep trying to move past him and Stirner, but have not managed so far.

P4 and eBPF are not really competing in the same space. eBPF/XDP is a virtual kernel machine that you can program in restricted C. P4 is a language that specifies how a device should process packets. So P4 is compiled down into a target (for example, a switch or NIC), which then executes the code as specified by the programmer. The nice part about P4, compared to just working with the devices, is that it establishes a standard reference format for administration and control of packet processing. For example, Google is using P4 to test their switches (they talk about it in their keynote in this workshop [0]).

The eBPF/XDP VM could be one of these P4 targets. You can write a P4 specification that is converted into eBPF byte code which is then loaded and executed. The reference P4 compiler even has a eBPF[1] and an XDP[2] back end. However, those are currently not well maintained. Another problem is that the code generated from P4 is not as efficient as manually crafted eBPF byte code. With a little care this is all easily solvable.

Disclaimer: I work in this space.

[0] https://opennetworking.org/2021-p4-workshop-content/

[1] https://github.com/p4lang/p4c/tree/main/backends/ebpf

[2] https://github.com/vmware/p4c-xdp

To explore Z3 projects, I sometimes just browse the tags on Github [0]. I also am not aware of a better way to get an overview. However, I know Z3 is used a lot in industry for verification and resource allocation tasks.

Personally, I have made extensive use of Z3 to find bugs in domain-specific language (DSL) compilers [1]. You can use Z3 to encode the source code before and after a compiler optimization as a logic formula and then check whether the formulae are equivalent. If they are not, there is likely a semantic bug in the transformation pass of your compiler, meaning you have introduced a subtle logic mistake.

This style of verification really works best if your programming language is sufficiently restricted (not Turing-complete) or if you pick isolated segments in your code (e.g., functions). If that is the case, these methods can be extremely effective. We have found many semantic bugs in a short time period. In our case, Z3 was not even the bottleneck, performance-wise. It was our language interpreter. And because the language is sufficiently restricted, we can even run our tool as part of the compiler CI. So whenever a pull request is made, we can check whether it may lead to subtle issues in program translation.

Another useful aspect of using Z3 to represent a programming language is that you can also generate test cases for your programs. As far as I know, the Rosette [2] framework is intended to make all of this more approachable.

[0]https://github.com/search?q=topic%3Az3&type=Repositories [1]https://www.usenix.org/conference/osdi20/presentation/ruffy [2]https://emina.github.io/rosette/

This is the most comprehensive discussion on this topic, which also deals with the fallout on Twitter. Includes (anonymous) comments by Google employees:

https://old.reddit.com/r/MachineLearning/comments/k77sxz/d_t...

Otherwise this search will cover most of the HN debate: https://hn.algolia.com/?q=gebru