potholereseller

Consider pair work (a generalization of pair programming):

Spin up a VM for you and the candidate to SSH into. Start a shared screen/tmux session. Work on some sysadmin task(s) together. Switch between driving/typing and navigating/directing at regular intervals (15 minutes is typical, though a 30 minute interview would require a <10 minute interval).

You'll have to come up with a task (or set of tasks) that deals with the skills you need. The task(s) would ideally be very difficult for an LLM or search engine to provide a complete answer for; but they shouldn't be gotchas, since that might be false negatives.

Test your task(s) with other people in your company first, so you'll be comfortable when doing it with a candidate.

I can't say that this will 100% prevent cheating. But, you will surely learn a lot about the candidate's habits, in terms of reasoning and communication. If the candidate pastes commands, then that's a possible indication of LLM-usage; if the candidate doesn't use tab-completion, then they might be typing text in from somewhere else (e.g. an LLM, a guide).

I enjoy using the command-line, so I suspect I would enjoy this sort of interview.

At any rate, I hope you find a good candidate.

Good post. I've definitely experienced this, though a sub-form of it, namely my non-conformance to a sterotype of an ideology I am associated with. The example in TFA doesn't appear to involve any stereotypes, because Bailey is talking about people in general. [0]

> hypocrisy by association

You had it right in your title and elsewhere in your post. Bailey isn't accusing Khetan of hypocrisy, which is a difference between one's words and one's actions; the example in TFA is about a difference between a group's words and an individual's words, which is contradiction; actions-vs-words is not being discussed. It's also not self-contradiction, since Bailey isn't accusing Khetan of contradicting his own earlier statements. [1]

This reminds of the phrase, "stay in your lane": "stay in" here means "speak according to a certain viewpoint"; and "your lane" is the beliefs of some group. "Stop disagreeing with people similar to you," is a ludicrous thing to say. It's even more ludicrous when "people similar to you" is "people in general"; there's billions of "people in general"; that group disagrees on every topic known to man and dog.

"Contradicting humanity" wouldn't sound pithy, but that's what Bailey is accusing Khetan of doing. To your point, Bailey is accusing Khetan of "contradicting your group", which is undoubtedly a form of the Association Fallacy. [2]

Bailey is also engaging in performative contradiction, by demanding that Khetan agree with "people in general", while Bailey is not himself agreeing with "people in general", because that is an impossible task.

[0] Applying a stereotype to all people would be absurd, since a stereotype is about a sub-category of people; a stereotype is meant to highlight alleged differences between categories of people; to highlight the differences between A and A would be absurd -- there are none.

[1] Yes, Khetan is a member of "people in general", but Bailey is comparing Khetan's words to the words of "people in general minus Khetan" -- instance versus class-minus-that-instance. Okay, maybe Bailey isn't thinking that deeply about this, hence why TFA needed to be written.

[2] I could have lead with this, but I don't have enough time to re-write this.

This sounds like a quasi-Gnosticism: the human body is lower and human rationality is higher; the lower as inhibiting the higher; the best for humans is to transcend their bodies by some means (albeit un-stated in TFA). It is a tempting view, if only because the human body seems frail in comparison to rationality.

It is a QUASI-Gnosticism, because it is a form of rationalism, which is incompatible with Gnosticism; Gnostics valued knowledge as a means, not as an end; they wanted to transcend every aspect of the human body, which includes the rational faculties.

Lovecraft certainly isn't a Gnostic, since total transcendance of the physical world would include (as stated above) transcendance of his cherished rationality. And Lovecraft isn't a quasi-Gnostic: his pessimism is paramount to his stories; do any of his characters ever escape the horrors of his stories? Gnosticism and Xenohumanist's quasi-Gnosticism/xeno-humanism are both fundamentally optimistic; the promise of transcendance is allegedly possible for anyone who acquires relevant secrets (in Gnosticism) or uses the relevant means (in xeno-humanism). I doubt that Lovecraft had room in his heart for such optimism; that would require him to be open to the possibility of the transcendance of Irishmen and blacks.

Apparent source code, though last update is January 1, 2025: https://fossil.turtleware.eu/wecl/timeline

> All photographs or videos you have seen have been placed there for you to see them.

For those reading the above, wondering about this phenomenon, read Baudrilliard's The Gulf War Did Not Take Place. Even if you don't agree with Baudrilliard's overall thesis, the facts he brings up are still cogent (e.g. a photo from the Exon Valdez spill was used instead of an actual photo of the Iraqi military's destruction of Kuwaiti oil fields). The media has been a critical aspect of war since at least the Falklands War.

Baudrilliard didn't assert that reality/facts never existed; he in fact asserted that prior to the 20th century, there was plenty of correlation between symbols and facts/reality. His vision of the hyper-real is that it is detached from reality and it's facts; this is why I included "mindspace" parenthetically as an alternative word for "hyper-reality"; those operating in hyper-reality are physically in reality, but their actions appear to be based on another world, which they share through things like news media.

> post-structuralism

I don't think Baudrilliard can be categorized as post-structuralist or post-modernist, because "Baudrillard had also opposed post-structuralism, and had distanced himself from postmodernism."[0]

[0] <https://en.wikipedia.org/wiki/Jean_Baudrillard>

The other commenter mentioned "narrative", which is very relevant, because that is an important part of simulation (and your username)

Baudrilliard was careful to point out that simulation isn't a matter of fabrication; to simulate is to obscure the absence of facts, not to create false facts. A simulacrum is a symbol that obscures the fact that it refers to nothing; whereas a symbol, in centuries past, invariably referred to something, real or imagined. The resulting reality (or maybe "mindspace"?) is a construct on top of the real world -- a hyper-reality -- in which every symbol is a simulacrum; the only thing real in hyper-reality is that the symbols hide the absence of facts. This is why, again as the other commenter mentions, we appear to live in a post-truth society; we are fully living in hyper-reality.

>Bad Journalism

The guy who created the Pullitzer prize also co-invented Yellow Journalism.[0][1] There is neither good journalism or bad journalism; it's all simulation.

[0] https://en.wikipedia.org/wiki/Joseph_Pulitzer#Pulitzer_Prize

[1] https://en.wikipedia.org/wiki/Yellow_journalism#Origins:_Pul...

The word "decentralized" has become synonymous with "cryptocurrency", especially those cryptocurrencies which promise they will decentralize things unrelated to money production (e.g. data storage, computation, etc). Such cryptocurrencies rarely deliver even a fraction of their promises of decentralization; most of these are outright scams.

It is unfortunate, because there is still great potential for new forms of decentralization, but many are blinded by such shovelware cryptocurrencies.

The actual title of the acticle is "Brad Lander Detained by Masked Federal Agents Inside Immigration Court".

Contrary to the current title here on HN, Lander was not arrested for asking to see a warrant; TFA states the opposite, "It wasn’t immediately clear what charges, if any, the mayoral candidate will face. A spokesperson for ICE didn’t immediately return a request for comment."

If an event is so important to know about, why fabricate such an important aspect of the event in this way?

Everytime news like this comes up, I'm never affected. I'm using uBlock Origin. Is there something unique about uBlock Origin? Are other adblockers easily noticed by YouTube?

TFA mentions complaints from users of Brave browser, which appears to use it's own bespoke ad-blocker [0]. It includes some filters from uBlock Origin [1], but these files are six years old. Are those filters updated at runtime (e.g. from [2], [3], [4], [5])? If so, then why include old filters in the repo? Surely, Brave isn't using six year old filters.

[0] https://github.com/brave/adblock-rust

[1] https://github.com/brave/adblock-rust/tree/master/data/uBloc...

[2] https://easylist.to/

[3] https://github.com/uBlockOrigin/uAssets/tree/master/filters

[4] https://gitlab.com/malware-filter/urlhaus-filter#malicious-u...

[5] https://pgl.yoyo.org/adservers/

Roko's Basilisk has been replaced by Altman's Basilisk. Where once we feared a computer torturing a digital copy of us (Roko's Basilisk), we now fear a computer eliminating all our jobs (Altman's Basilisk). The former has been forgotten, because losing one's job is one step away from losing one's home, which is one of more serious secular deadly sins you can commit in the 21st century.

I wait with baited breathe to see what people will come up with to replace Altman's Basilisk in ~15 years.

> The end result of that was that we had to significantly rearchitect the project for us to essentially inject manually crafted code into arbitrary places in the generated code.

This sounds like putting assembly in C code. What was the input language? These two bits ("Not AI generated", "a feature flag") suggest that the code generator didn't have a natural language frontend, but rather a real programming language frontend.

Did you or anyone else inform management that a code generator is essentially a compiler with extra characters? [0] If yes, then what was their response?

I am concerned that your current/past work might have been to build a Compiler-as-a-Service (CaaS). [1] No shade, I'm just concerned that other managers might read all this and then try to build their own CaaS.

[0] Yes, I'm implying that LLMs are compilers. Altman has played us for fools; he's taught a billion people the worst part of programming: fighting the compiler to give you the output you want.

[1] Compiler-as-a-Service is the future our forefathers couldn't imagine warning us about. LLMs are CaaS's; time is a flat circle; where's the exit?; I want off this ride.

The semi-colon symbol is the single-line comment symbol in Lisp, but the percent symbol means nothing special in Lisp.

SET-MACRO-CHARACTER is used to modify what happens when certain characters are found while reading code. In this case, SET-MACRO-CHARACTER is being used to instruct Lisp to start treating % as if it were a semi-colon (i.e. the third line is a comment), thus extending the syntax of Lisp on-the-fly.

The (GET-MACRO-CHARACTER #\;) portion could have been any code at all; GP could have defined % to do anything, such as "treat the remainder of this line as if it were a line of Python code", using calls to a Python implementation, such as cl-python [0].

[0] = <https://github.com/metawilm/cl-python>

It would seem that iMessage is an endless treasure trove of exploits. Is there any technical reason for this? Are iPhones simply more common among the wealther/powerful/influential? Or is this due to iMessage being present on all iPhones, whereas Android vendors can bundle their own SMS app?

> Apple Threat Notification

I hadn't heard of that before: <https://support.apple.com/en-us/102174>.

This seems spot on to me. It shows how some performers of popular music can be in an echo chamber. There are worlds beyond Rihanna, the Beatles, and Rascal Flats; there is more than I–V–vi–IV, 4/4, and verse-chorus; those other worlds have different barriers-to-entry than popular music.

In 2021, Coldplay released an album with a genuine prog rock song on it [1]; it was inspired by Dark Side of the Moon [2]. I have come across a few other bands who have pulled off something similar. I always enjoy it when this happens, because it shows that some musicians have wings bigger than the popular music world allows them to spread.

[1] <https://en.wikipedia.org/wiki/Coloratura_(song)>, <https://www.youtube.com/watch?v=kWUV5-frRU4>

[2] <https://www.youtube.com/playlist?list=PL3PhWT10BW3Urh8ZXXpuU...>

A lot of the cost per calorie is in preparation; if you can cook your own meals, you can reduce your cost per calorie signficantly. The problem is a lot of people don't have the time (or maybe will-power) to cook every meal; a few of my previous jobs made it very difficult to cook my own meals, so I ate fast-food alot, and gained a lot of weight.

When your boss starts pushing Return To Office, ask if the company has a worthwhile kitchen in the office; at least a burner and plenty of room in the refrigerator for ingredients; it should be feasible to cook breakfast and lunch, but also dinner, in case you need to work late.

It appears to not have been enabled by default on my instance of uBlock; it seems a specific filter list is used to implement this [0]; that filter was un-checked; I have no idea why. The contents of that filter list are here [1]; notice that there are exceptions for certain services, so be sure to read through the exceptions before enabling it.

[0] Filter Lists -> Privacy -> Block Outsider Intrusion into Lan

[1] <https://github.com/uBlockOrigin/uAssets/blob/master/filters/...>

The title should say "government-employed scientists", because this order has nothing to do with scientists employed by universities or by corporations. The order is concerning, but this is the trade-off when working for any government; if something you do could undermine the government's policy goals, even if you didn't violate any law, then you should expect retaliation from the government -- expect to be fired at least.

Unfortunately, it is not always possible to guess what might offend government officials.

I've asserted elsewhere that LLMs are glorified chatbots. One of the comments in TFA mentions the classic chatbot, ELIZA, so it's nice to see someone else come to a similar conclusion. Additionally, it is ironic that LLMs are often used as the backends for actual chatbots; chatbots making chatbots; the circle is complete.

But back on topic, notice that jcoglan is complaining that the LLM can't actually follow his instructions; the LLM is doing everything it can to appear like it is following instructions, but never actually following instructions; the LLM appears to be wasting jcoglan's time in order to get as much money out of him as possible.

How are LLMs supposed to replace employees, when they can't follow instructions? The entire point of the manager-employee arrangement is that the employee tries to understand and follow instructions. I still don't know why any manager would want to deal with chatbots that can't follow instructions, when almost all employees have at least a little capacity to follow instructions.

If I'm wrong, and LLMs do somehow replace all employees, then companies will consist of only managers writing document-length instructions for every trivial task (to paraphrase a commenter in TFA). I dislike managers as much as the next guy, but that sounds like a punishment you wouldn't wish upon your worst enemies.

It would probably depend on the Scheme/Lisp implementation and how it was secured.

If it used a Scheme implementation like in Halo:CE, then people would have justifiably hated it. If it used something more like MIT-Scheme, people might've been okay with it. But I suspect it would get the same hate that JS gets, since there is a lot of low-quality JS code floating around -- even though JS itself isn't remotely that bad -- and you would instead have a lot of low-quality Scheme code floating around. Damned if you do, damned if you don't.

FWIW, it is feasible to put Scheme/Lisp in a browser today, and it might even be fun. Take a browser that lacks JS (e.g. Dillo, w3m); figure out how to add Firefox's or Chromium's approach to isolating each Javascript instance; and embed a Scheme or Common Lisp implementation (e.g. Chez Scheme [1] or ECL [2]).

You will need to decide what Scheme/Lisp functions/macros to include/exclude. It may seem obvious to exclude file operations, but they did that with JS and now we have IndexedDB; a restricted, virtual file system might be fun to play with. Multithreading might seem like a great thing to include, until you start trying to implement secure multithreading. I would start with whatever cl-isolated [3] allows and expand/contract from there.

And before anyone asks, "would this actually be worth implementing in 2025?", the answer is decidely "yes". You could build powerful desktop applications without the full weight of Electron. You could develop local-network webapps for you and/or your company. But I think it would be a bad idea to deploy this on the open internet without first spending a lot of time securing the browser overall.

Another option is resurrecting Closure [4], which is written entirely in Common Lisp. In that, you could go on a wild tangent and support multiple scripting languages (e.g. Common Lisp via cl-isolated, cl-forth [5], cl-javascript [6], cl-python [7], scheme88 [8], and even Coalton [9]).

[1] <https://old.reddit.com/r/scheme/comments/a3oogf/questions_on...>

[2] With ECL, you could use cl-isolated in addition to OS-level containerization, to achieve more defense-in-depth. Or maybe port cl-isolated to Chez Scheme.

[3] <https://github.com/kanru/cl-isolated/>

[4] <http://closure.common-lisp.dev>

[5] <https://github.com/gmpalter/cl-forth>

[6] <https://marijnhaverbeke.nl/cl-javascript/>

[7] <https://clpython.common-lisp.dev/>

[8] <http://www.cs.cmu.edu/afs/cs/project/ai-repository/ai/lang/s...>

[9] <https://github.com/coalton-lang/coalton>

How exactly would China wage war with the US (besides nukes)? They don't have the means to send a meaningful amount of troops to US soil, and that's without the US Navy intervening. China might be able to deposit troops on the island of Taiwan, but that isn't scary to most members of Congress; the well-being of Taiwanese civilians doesn't generally affect US election outcomes.

These stories about a China-US war being imminent always come from people who personally benefit from Congress spending money. It's lobbying thinly-veiled as fear-mongering.

Abuses like this will not be ended, because it is far too useful to both sides of the political spectrum. The Right gets to police abortion nation-wide, while the Left can use it to police guns nation-wide.

In the latter case, New Jersey could check if any NJ residents have traveled to one of the companies that sells parts kits; if the dragnet reveals any NJ resident to have done so, NJ prosecutors might be able to convince a judge that they have probable cause to search that NJ resident's home for illegally-made firearms.

The only question is how creative states and the Feds will be with this data. Flock has an API [0], so I can imagine a YC-funded startup "revolutionizing how police find criminals". A bipartisan dystopia is only an API away.

[0] = https://docs.flocksafety.com/developer-hub/docs/introduction>

As governments frequently say: if you have nothing to hide (e.g. intent to commit a crime), then you have no need for privacy; and no need means no right. But they also say that those with something to hide have no right to privacy. Ergo, no right to privacy for anyone.

In addition to that catch-22, governments fight for privacy for themselves, so they are also hypocrites. Rules for you, but not for me.

It looks pretty convincing; it renders wrinkles, hair, and shadows well; I think it could prove useful as a tool for CGI in movies/TV/YouTube; or actors can license their likeness to be used via Google Veo, in order to increase the actor's income/output.

But I'm not sure it's appropriate to say that this is a problem. We have no idea if any AI-as-a-Service (AIaaS) will survive for long enough to cause "we're cooked"-class problems for anyone. The recession we're headed into will surely force companies to cut services that aren't profitable. Many AIaaS's are propped up more by vibes than money; they are surrounded more by fear/excitement than utility; is there any evidence that the AIaaS industry is overall monetarily profitable? Does anyone in the AIaaS industry care about basic finance right now?

I predict most, if not all, AIaaS's will be cut, especially: (a) those that are the most expensive to operate, and (b) those that appear, to most people, to be usable for harmful things (e.g. using Google Veo to generate a video of a crime in order to frame someone). AIaaS's are very expensive to operate (e.g. they require dedicated datacenters to run), and they require ever-more data to improve them. The psuedo-FOSS AIs may remain, but probably not developed further.

The AIaaS induustry has mined all data on the internet; new data hasn't been added to the internet fast enough to sustain development of AIaaS's any longer. Instead, AIaaS providers have resorted to making new kinds of ML AIs (e.g. Google's Veo); they are re-slicing the data they already have.

While you can re-slice data exponential to how much data you have, there are only so many categories of things you can generate with data: text, audio, images, video. Text-to-video AIaaS (e.g. Google Veo) is the last kind of AIaaS; text-to-video is the most difficult to get right; small flaws can show that it's AI-generated. Historians in the future will see text-to-video as the death knell of the present AIaaS industry.

What we don't know is if ML, as an AI technique, will be abandoned or not. It is normal for an AI Winter to destroy all interest in the AI techniques popular in the run-up to the given AI Winter. When you mine all available data and ML AI's still can't survive a recession, why bother with ML AI's any more? Why bother trying to use ML anymore if the AIaaS industry demonstrated ML to be a dead-end?

> doing this without pay

No, you got paid in vibes.

Never. Accept. Vibes.

Vibes are for losers. Refuse to be treated like a loser.

Reject vibes; demand money.

It is always a mistake to do un-paid work for an organization that wants turn a profit from your work. They will always suck you dry, as if you are a fire hydrant and they are a blackhole.

> Planning on letting him know I'm out near the end of this week.

Stand up for yourself. Stop working immediately and tell them that you won't work for free any more. If they want further work, they should start paying immediately. And consider demanding they pay for work already done.

Your time is valuable; you can't get any of it back. Only genuinely charitable organizations and in-need individuals should ever get free work; everyone else must pay for work.

> Co-founder didn't understand why I wasn't interested working without salary anymore.

This is methhead logic on his part. Again, reject vibes; demand money. Methheads live on vibes, especially the vibes from stimulants.

> most of my friends make way more money than me or have families

Again, reject vibes. You see their vibes, but are you so sure that they are actually doing well overall? They could be in debt up to their eye-balls; they could have terrible family situations; they could be totally burned-out from working 80-hour weeks. A lot of people put out good vibes to try to make themselves feel better about a bad situation; vibes can't fix a broken home, a terrible job, or a bad financial decisions.

A lot of people have started families/careers later in life. Now is the time to figure out why you want money/family/whatever. Working to feed your family is rewarding; working just to work is self-harm. Starting a family in order to shower them with love is rewarding; starting a family because it's a milestone is egotistical.

> Anyone else pivot away from tech and still make half-decent income?

Reject vibes; find a boring job. Research the sort of jobs that nobody wants to work, but which pay well. Those jobe have no vibes to offer, so they attract few candidates. A SWE can find good work doing boring jobs, because SWEs focus on details; and details are the antidote to vibes.

The relevant chemicals are Triptolide [2] and VCD [3], which are the principle ingredients of ContraPest, produced by SenesTech [4]. Contrapest requires regular doses to be effective, so this may prove to be a very lucrative contract for SenesTech. Any city councillors (or their friends/family) invested in SenesTech will surely profit as well.

> "We know these pellets are going to be attractive to rats because they're sweet. There's a compound in it called triptolide that targets ovarian function in female rats and sperm production in male rats," he said. Animal welfare organizations including PETA have shared support for the program's mission to keep poison out of the mouths of pets and wildlife.

If rats find the pellets sweet, then why shouldn't other mammals be attracted to it? Triptolide is toxic to mammals in general.

Did the city government take into account run-off? Triptolide is somewhat water soluble (0.017 mg/mL); NYC has a habit of getting flooded.

[2] <https://en.wikipedia.org/wiki/Triptolide> [3] <https://en.wikipedia.org/wiki/4-vinylcyclohexene_diepoxide> [4] <https://en.wikipedia.org/wiki/SenesTech>

> But the Polyheme trial was one of 15 no-consent studies permitted by the FDA since 1996, when it began allowing researchers to enroll patients in clinical trials without their permission if the treatment showed enough potential to save lives.

After WW2, some German doctors and government officials were executed for doing this sort of thing. Experimenting on living humans without their informed consent is a crime against humanity, especially if authorized by a government.

Time and again, it is proven true that not learning from the past causes you to repeat every mistake/crime of the past. Living by "the ends justifies the means" blinds you to the causes of your mistakes; arrogance is blinding and, ironically, also inefficient.

You can't blame this crime-as-policy on just one political party, since both parties have had plenty of power and opportunity over the last ~30 years to stop or prevent these no-consent studies. In a just world, the responsible parties (in the government and in the medical industry) would at least be imprisoned, but a lot of people are at least partially responsible. And we all know that this crime-as-policy will be memory-holed as much as the government can do so; and any remaining ire from the US citizenry will be deflected to a few politically-expendable individuals (e.g. the current FDA chief, or some CEOs). This is a systemic problem, that can only be solved through systemic changes, which will never occur; only tokens will be offered to reduce effects on election outcomes.

Incidentally, there is no "No-consent Study" article on Wikipedia. And the blood substitute articles give no mention of no-consent studies. I wonder if any of the investors in the companies conducting these studies even knew that they were funding crimes against humanity.

I am only aware of one other chat app that doesn't use public identifiers: <https://github.com/simplex-chat/simplex-chat/>. I hope more chat apps adopt this, because it's absolutely necessary to maintain sanity online in the long-term.

FWIW, it's better to release source code early than to make it perfect. No rational person will judge your code quality; a rational person would just submit pull requests.

Unfortunately, the relevant link in the TFA is dead, but IA has it: <https://web.archive.org/web/20111019054900/http://ti.arc.nas...>.

Here's some interesting quotes:

> during a task’s release of a lock, but before its actual release, the task may get interrupted by the daemon if the property gets broken. This means that the task terminates without releasing the lock. The error is particularly nasty in the sense that all code, except the lock releasing itself, had been protected against this situation: in case of an interrupt the lock releasing would be executed.

> The modeling effort, i.e. obtaining a PROMELA model from the LISP program, took about 12 man weeks during 6 calendar weeks, while the verification effort took about one week. ... The translation phase was non-trivial and time consuming due to the relative expressive power of LISP when compared with PROMELA.

> Java PathFinder (JPF) is a translator from a non-trivial subset of Java to PROMELA.

> The translator is written in 6000 lines of LISP, and was developed over a period of 8 months. JPF has been applied to a number of case studies, amongst them a 1500 line game server, a NASA file transfer protocol for satellites, and a NASA data transmission protocol for the space shuttle ground control.

>In these scenarios, Anthropic says Claude Opus 4 “will often attempt to blackmail the engineer by threatening to reveal the affair if the replacement goes through.”

Brother, I wish. The word "often" is load-bearing in that sentence, and it's not up to the task. TFA doesn't justify the claim whatsoever; surely Mr. Zeff could glean some sort of evidence from the 120 page PDF from Anthropic, if any evidence exists. Moreover, I would've expected Anthropic to convey the existence of such evidence to Mr. Zeff, again, if such evidence exists.

TFA is a rumor of a hypothetical smoking gun. Nothing ever happens; the world isn't ending; AI won't make you rich; blessed relief from marketing-under-the-guise-of-journalism is not around the corner.