Postman was my curl with a gui. Used it everywhere, recommended it everywhere. But one day out of the blue it asks me to login, and it all went downhill from there. Imagine you had to login to use curl?
Anyway, I'm sure somewhere in some terms and conditions I agreed to, it says "we own the glass".
This thread reminds me of the time I was trying to learn meteorjs and people ask me if it has html5 shims. I google it, then I get asked how does it handle state, and if it supports functional, i google... Does it hydrate the dom? What about shadow dom? Let me goo... Does it have hooks?
Anyway, I just learned what abliterated means and it might be obsolete by tomorrow.
Edit:
> An abliterated large language model (LLM) is an open-weight model modified to remove or dampen its built-in safety refusals without requiring full retraining or prompt jailbreaks.
It's a patched model with less censorship/refusal.
> An abliterated model is an open-weight artificial intelligence (Al) language model that has been modified to remove or reduce its built-in refusal behavior without retraining the entire neural network.
Since no one gave you a real world example of how this would be useful - Qwen 3.8 27B runs locally for me and is more than capable of controlling my sonarr and radarr stuff for downloading and managing and organizing my media library but will flatly refuse to do any of that. In fact it’ll get very snarky with me. An abliterated model would have no such hangups.
On my side, I built a full sonarr radarr qbitorrent tracker and znb solution on a nas only with Claude code. He help on each step, make install and maintenance script. He never said no.
TLDR "uncensored" so you ask it stuff the vanilla model would refuse and most likely it would comply/answer.
I'd advice caution with abliterated models for various reasons. there was even a couple of days (or weeks I forgot) article here on ynews about how some of these models can get you pawned (https://projectdiscovery.io/research/how-abliterated-models-...)
Not to brag*, but that's what I've been doing since 2013 on my php blog [0]. It was actually the result of a hn warm hug of death. My CDN is just a separate webserver serving all static assets using nginx.
Both my css and js are dynamically minified and included in the html depending on which component is loaded on the page.
With php2shell going mainstream, to the point that random friends who happen to know I'm a programmer, consult me to understand the extent of it, I think they succeeded with their doomerism message.
When we find a bug, we patch it. When we find more bugs, we patch them as well. That's what happened here. Sam is telling us, "we found bugs, it's over" and the general public swallowed the frog.
I mean, there is the bbc interview where they invited a cyber security expert who said he didn't know php existed before the the news broke out. The general public doesn't see that gaff.
But you already have Claude-directed videos that utilize OP's manim project (which he uses himself to do 3b1b) that anyone else can create. Grant open-sourced manim and now anyone can point a model at it to create their own 3b1b-like videos. They're popping up all over the place as Opus 5.5 demo videos just in the last few weeks. So... AGI?
I guess I wasn’t clear, but using manim to make videos is not the hard part. The hard part is making them engaging and enlightening, and whatever else you want to attribute to 3b1b videos which make them a joy to watch and learn from.
I tried watching 3b1b over the years, really wanted to enjoy it. But I never found the way he presents things intuitive and never got a palpable intuition that helped me with math classes or work.
I saw some clearly AI slop made videos and got much better understanding of topics, since they don't do foreshadowing for an awe-inspiring connection that we'd admire together after 15 min videos, the clankers just get to the point. Not claiming those videos are too useful either, but I did however see accelerated learning gains from asking LLMs to explain (and do Bret Victor-style small interactive demos) math and algos, something I have never experienced from any of the educational STEM Youtube channels.
Even if you find them intuitive, they produce no learning gains for anyone. Take a lin alg test, watch the 'Essence of lin alg' videos, take comparable test again. I guarantee you won't see any change even if you 'understood matrices better'. They are pure mathporn.
No learning gains? I teach linear algebra. I can assure you that students who watch and study the "Essence of linear algebra" videos (which I recommend on the syllabus) do better on my tests that those who don't, because they learn stuff from them.
It's much more likely that the students who take the time to watch supplementary materials are the ones who'd do better regardless of the quality of the given material. If you have access to fresh learners, I'd say it's a good opportunity to test how good those series are.
I'm not sure anything I say can convince you, as your mind seems to be made up already, but my experience has been that those videos help my students understand the concepts of linear algebra, comparing the mastery level now vs. before the series was published on YouTube. Especially when it comes to getting a feeling for what a linear transformation is, and how the properties of matrices echo the properties of the transformation they represent.
As for "fresh learners", I can provide my own son, who watched the series before taking his first linear algebra class, and says that it helped him a lot.
I'm not defining education as passing tests. A test is just a measuring instrument.
In fact, your claim is strong and quite easy to test. If watching the series gives people a better fundamental understanding which makes subsequent learning more efficient, take matched novices, randomise them to watch the series or not, then teach both groups some later lin alg materials and measure learning gains on novel problems.
That's the kind of test I mean, not 'can you grind through the course exam'.
If your claimed deeper understanding neither improves what someone can do afterwards nor how efficiently they learn related material, I'm not sure what observable content the claim has.
This is not my experience. Having an intuitive grasp of e.g. the graphical interpretation of the determinant instantly lets you figure out most of the key determinant properties without having to memorize them. That's the primary thing you get from intuition imo, having a sense of what must be true without having to memorize or prove it.
I'm actually quite optimistic, was pointing out that, at least to to me, AI slop already produces better results than big educational content creators. For something concrete, even if not classic math, I credit direct LLM interaction highly for accelerating my understanding of CPU design and compiler tech.
fwiw the fact that brains work fundamentally differently was my big penny dropping realization after digging into what LLMs are doing. Everything (I assert) a brain does (the function between inputs and outputs) is due to training data. There is no "instinct" and there is no "ROM-driven way to think" in any brain. It's training data all the way down. As a consequence different brains do indeed think in fundamentally different ways.
This is so fundamentally divorced from what we know about cognition in humans or in animals. The idea that we are a pure tabula rasa and only shaped by our "training data" is absurd and false. Just look at the amount of data we need to be able to do things relative to ML models. Look at animals which can do complex tasks moments after birth. We're deeply, deeply shaped by instinct.
Well, everyone being tabula rasa has a point tho. Technically there's no unique physical cause that make development unfold the same for all organisms, or even all humans (see Waddington's epigenetic landscape). It just happens that we develop along very similar trajectories and can draw equivalences between how I think vs how you think. So, from an ontological perspective, if you squint a bit, we all are tabula rasa. But from a phenotypic perspective, of course it's bonkers, you can't teach a frog linear algebra.
Last year, I wrote a post about how I use zipbombs to deal with bot traffic [0]. It went mega viral over the web, and I regret it. My server was completely overwhelmed with bot traffic. After months of dealing with it, I had to find a better strategy then keep my mouth shut.
Last year, I read your post and implemented it. I had no way to tell whether there is any benefit to it, so I discontinued it. Serving the data, even though it is compressed and the compression is precomputed, was still a waste of bandwidth. It could hit the browsers of legit user (false positive).
The main problem on my site is that bots hit the CGIT repo, where visiting files triggers syntax highlighting scripts.
I put in a skill-testing-question type system where you get a maximally long-lived access cookie if you answer right (so legit users are bothered as infrequently as possible).
CGIT doesn't support client-side syntax highlighting, as far as I know. Anything could be hacked, but out of the box, it has a system consisting of dispatching a script.
I have custom syntax highlighting. For instance, I have a home-grown script which takes a man page file (identified by a suffix like .1) and renders it readably, so it looks like a man page on your terminal, rather than nroff source code. (The only unsolved problem is that CGIT renders a column of the original line numbers down the left, which exceeds the length of the rendered page.) I also have accurate syntax highlighting for a home-grown programming language. That syntax highlighting exists only for Vim, which has a superb syntax highlighting engine. So in fact Vim is launched, via expect, to do some HTML rendering! There is no way that is getting translated into client-side.
Pre-rendering is not really an option over numerous git repos, that have numerous files in them, with the added dimension of history. What CGIT has instead is tunable caching.
Bots which aggressively scrape the history are hostile to the caching scheme, however. It works best for human visitors which mostly look at the head of the master branch in each repo, and poke around just a little bit in the history or look at other branches.
It worked fine... on a small scale But it's one thing to swat flies versus fighting swarm of yellow jackets. Serving a 1MB file a million times a day on a small digitalocean droplet is not feasible. Let's just say it didn't scale. And AI bots usually come with beefier machines, and recover faster after a bomb.
It's the content of the page, the raw html, that is replaced with a zipbomb. Example: https://example.com checks if you are a malicious bot, depending on the result, it will send you gzip html or gzipped bomb. So that can't be swapped or redirected to a CDN. The bots would simply ignore the redirect.
I have a small website with a contact form that gets a lot of spam. I found a very effective way of filtering spam, but it only works because it's not mainstream, so I can't ever share it :(
Once I had an issue where I had to request access to a 3rd party vendor URL from my work machine. IT was telling me that I already had access, and that I had accessed the URL several times. Weird, because I was getting a dns resolve error. To prove that he wasn't making it up, the IT guy shared his screen with me.
Ever since I saw his screen and the level of details in my activity log on the work machine, I don't even access example.com on their device anymore.
I'm sure this workers already has invasive tracking software on her work machines. Now the company was just trying to fire her without it counting as a layoff.
If you're working remotely, the solution is to have a personal laptop next to your work laptop, and use the personal laptop for AI, email, apps etc., even if it is for work (unless you are using a service that costs money). Your employer should be fine with this arrangement because all they care about is that their data is not exfiltrated.
Depends on what counts “employer data”. If you type a question into ChatGPT is that “employer data” even if it is a work related coding question? Probably not.
If an employer is using Zscaler and employees can still “pass employer data” anyway then Zscaler isn’t achieving anything. Get rid of it so employees won’t try to work around such intrusive monitoring.
This seems to be a bizarre cross purposes discussion. If I'm typing a work related question into an AI then that is clearly employer data .. which I'm happy to type into the work laptop. It's the one with the corporate license! I can easily cut and paste into/out of other apps!
Where this gets a bit more complicated is doing stuff relates to the matter of employment on the work laptop. The classic example of this going badly is union organization, which in many but not all places is a legally protected activity but may still invite illegal retaliation. HR disputes are another one where it may be extremely necessary to "exfiltrate" some emails regarding your terms and conditions.
I imagine they meant the type of question that's generic on its own, but related to what they're doing at work. IOW, identifiers replaced with "foobar", relevant conditions isolated into brief and portable reproduction steps that don't reveal anything of substance about the company, etc.
If they allow you to access email, work apps, etc off corporate network then they don’t care and you’re not exfiltrating anything against their wishes. To suggest otherwise is just goofy. They weren’t suggesting you move data via a usb drive or something. Even that your employer would disable if they cared.
That is almost textbook data exfiltration, with company data on a personally-owned device. This is almost certainly forbidden by policy, and if it isn’t, it should be.
Do work stuff on work hardware, do non-work stuff on non-work hardware.
If you are asking ChatGPT a coding question does it matter if it is done on a work machine or a personal machine?
A company that monitors every keystroke using Zscaler can be presumed to have restrictions on stuff that they care about, such as apps that should only be accessed from the corporate network.
And that's precisely the benefit of using the personal machine! Nobody wants every keystroke they make to be monitored. If you do that then it is going to backfire because people will find work arounds. Find another way.
Or you can do a Show HN once you get fired for not understanding what most people are willing to put up with, thereby risking your employer's data that way!
You may be surprised to discover that many WFH/remote jobs are BYOD!
I held a WFH and BYOD job in various roles for over 4 years. It began less than a month into the pandemic lockdowns. It would not have been WFH if not for COVID-19. In that regard, I was highly blessed.
The recruiting process was notably vague. I was asked if I had a computer that I could run software on. It turned out that they really wanted us to be running Windows (Apple Silicon was incompatible) but I was running Ubuntu and nobody really batted an eye, because I ensured it was interoperable.
We were never required to install MDM or any corporate software. They never spied on us or collected any data from our BYOD. Any metrics were collected server-side at their web apps. I felt highly trusted. But there was indeed no "work laptop" and we could either BYOD, or forget having a job at all.
This was also during a period of rapid, widespread MFA adoption. Our managers were keeping us in the loop whenever MFA was implemented for corporate systems. Late in the game, our 3rd-party payroll implemented MFA very poorly. I informed the management that they were on thin ice, because they had irrevocably encroached onto a device which was reserved for personal use only. I told them, it is unacceptable to require a smartphone for a role that neither provides a smartphone, nor notified us in advance that it's required for keeping our jobs.
I kept scrupulous separation of work-related data and activities. I had separate accounts on each device. For work, I logged out of my personal account and logged into a work-only account. Unix permissions handled the rest. Sometimes, it was darned inconvenient.
I came to understand that if my employer were ever sued or involved in criminal prosecution, my devices could've been subject to subpoena and confiscation. This was the really crazy part. I had often used personal devices, and accounts, and my phone, for volunteering at church. We had no choice there, either. But you can easily imagine church, or a publicly-traded company, to become embroiled in some legal action. And if my devices were deemed evidence, I'd forfeit them, possibly for good! Thankfully, this has not come to pass.
There is probably no guarantee, even if you're 100% using employer-dedicated hardware at home, that your personal devices wouldn't be swept up in a subpoena or search-and-seizure. But you should always consider long and hard, with any BYOD situation, that mingling your personal stuff with your employer's stuff is playing with fire. Don't get burned.
Fortunately all my employment contracts thus far have been "the company shall provide all necessary equipment for work" and prohibition of using personal devices, which I have cited a few times when asked to do otherwise. Few people seem to understand the legal implications or downplay them, but IMHO it's best to keep a strict "firewall" between company and personal.
Before I retired (thank the gods) the company I worked for allowed byod if people didn't want to carry two phones, but it was made abundantly clear that their personal device was then open to discovery in case of any future lawsuits involving the company.
When COVID-19 hit, the only prep my employer had was that everyone had laptops and we had basic VPN infrastructure. The first week or two were chaotic. Everyone had to provide their own high-speed Internet. Eventually they let us come in small groups to the office to bring home monitors, keyboards, and mice.
Our company was never BYOD, but they were pretty loose about personal hardware. They didn't provide or require phones, but to get office mail, documents, etc. you had to have management software on your phone or tablet. You couldn't connect your own computer to the VPN.
We came back to work in mid-2021. In early 2022 we had a ransomware breach on a Sunday evening. If you happened to have your laptop with you (I did), you could eventually do some WFH stuff, after certain servers were certified clean (they never told us, but indications were someone succumbed to a phishing attack). Otherwise you were SOL until they either certified your machine had been turned off or was otherwise certified clean, or you got a brand new laptop.
Once we were back in the office (again), things were stricter...no USB drives allowed, for instance, and software installs were more tightly controlled. But the phone/tablet policy was the same -- even if you were on call, you had to provide your own phone. MFA could be on your phone or software on your laptop, and again you weren't required to have company software on your phone.
As I just got done writing, in many BYOD jobs, you can either use your personal device for work, or you can file for unemployment.
> The timing seems a little off
I shouldn't have to explain myself to people who get suspicious that I'm lying, but here's how it went down: our students were 100% required to have an x86 notebook that ran Windows, and we were explicit about that in the enrollment and onboarding phases.
Some of the staff used MacOS and this was not an obstacle, until Apple Silicon became available, at which point we discovered that our virtualization software was incompatible, and that couldn't be changed. We had been at the forefront of testing and vetting this platform on its release. Therefore, the department instituted a ban on the Apple Silicon devices.
My use of Ubuntu and Fedora, however, flew under the radar for all four years, because none of the software was incompatible. There were instances of staff, and also students, who purchased Apples specifically, and they were real disappointed, to say the least.
Ultimately, the staff in my role and department were relieved of the requirement to install any software locally, at which point anyone running Apple Silicon could've also flown under the radar. But our students were prohibited, due to the unresolved technical barriers.
Never, ever, ever ever use your personal laptop for work. Never use your personal mobile for work unless you have MDM software.
No employer with a functional legal department would ever be okay with an employee using their personal devices for work (unless they have MDM software on the devices partition the work data).
Using your personal device(s) work for make them fair game for discovery in lawsuits, even after you have left the company. It is also generally a violation of state and federal laws and/or procurement contracts in multiple fields if you work in finance, health, or anything defense-related.
If an employer ever mandates that you use your own device for mandatory work-related functions, find a new job. It's the number one red flag that the company is a shit-show and indicates that there are a lot of other rules/regulations/etc that the company doesn't care about either.
> No employer with a functional legal department would ever be okay with an employee using their personal devices for work (unless they have MDM software on the devices partition the work data).
This works the other way as well. While I (as an individual) have no legal departement, there is no way I'd ever be ok allowing work to touch anything on my devices, MDM or not.
My devices are mine, they will contain zero work stuff. Any device that workplace wants me to use for work, will be provided by the company, not me. I'll never allow any overlap.
(20 years ago I was at 100% overlap, no boundaries at all, but times have changed.)
I've been working on this industry for 3 decades. Work devices becoming standard to the point of seeming normal is one aspect of the industrialization of this job. I don't miss everything from the past era of craftsmanship, but I do miss working in a high trust environment.
I expect a lot of people are using Microsoft apps which do their sandboxing at the app level, coordinated with Authenticator, and not MDM. Ends up with the same level of control of company data but no device level control.
> > Do all your work activity on work devices. Do all your personal activity on personal devices.
> Not when a company monitors every keystroke.
That is precisely why you do work activity on work devices and anything personal on things you own.
I used to mix all the time but in these days of completely monitoring of work devices, I do strictly all work activity on work devices. Then I don't really care about the monitoring because anything I type on a work devices is strictly work stuff, nothing that is personal to me.
At my job, it was encouraged to to enroll your personal phone with the company's configs to make it easier to do work things. After a few months I took it off, it creeped me out and wasn't necessary to do my job.
I am not necessarily opposed to putting things like messaging or email on my personal phone but I’m not going as far as allowing the company to administer my personal device.
Good advice. But most work computers I’ve used were handed to me fresh from the Apple Store and they requested I wipe them before returning at the end of my employment. Not everyone works in places with an IT department, device management systems, etc.
While parent may have indeed been provided an ordinary stock Apple device, Apple's "Automated Device Enrollment" program can be used to administer/manage devices even when shipped directly from Apple to employees. I got a laptop replaced at a former company, shipped directly from Apple to me. I expected to have to reach out to IT to get it bootstrapped with their management tools, but the OS install process did this automatically.
> Not everyone works in places with an IT department
True, if you work for very early startups with no IT department, life is a lot more free. Just be careful when they grow to the point of having that IT department.
Also be aware of peer comment, devices can be shipped direct from manufacturer but already pre-registered to be company-managed. But this does require said IT department to exist to be managing this.
I do make exception to my hard rule on not sharing company/personal activities on devices when working with tiny startups where this is not an issue.
https://web.archive.org/web/20261108193851/https://sergeykuh...
reply