With this and other projects like Hyprland I do wonder what the right thing to do is. Like, the project would probably be better served if the governance was not dickheads. But they make decent products for free. The community could fork with a better governance model and just continuously merge the work from the “problematic” people involved, but that seems like a lot of work for very little gain.

At the same time I would not want to get very involved with community development efforts led by people who do not share my core values, and I’m probably not alone, so we’re probably missing out on opportunities by keeping the status quo.

I don’t think that’s really fair. There are cranky contradictarians everywhere, but in my experience that feature has been well received even in the AI-skeptic tech circles that are well educated on the matter.

Besides, the technical “concerns” are only the tip of the iceberg. The reality is that people complaining about AI often fall back to those concerns because they can’t articulate how most AI fucking sucks to use. It’s an eldtritch version of clippy. It’s inhuman and creepy in an uncanny valley kind of way, half the time it doesn’t even fucking work right and even if it does it’s less efficient than having a competent person (usually me) do the work.

Auto translation or live transcription tools are narrowly-focused tools that just work, don’t get in the way, and don’t try to get me to talk to them like they are a person. Who cares whether it’s an LLM. What matters is that it’s a completely different vibe. It’s useful, out of my way when I don’t need it, and isn’t pretending to have a first name. That’s what I want from my computer. And I haven’t seen significant backlash to that sentiment even in very left-wing tech circles.

Honestly, PCs are not ready for local LLMs

The auto-translation LLM runs locally and works fine. Not quite as good as deepl but perfectly competent. That’s the one “AI” feature which is largely uncontroversial because it’s actually useful, unobtrusive, and privacy-enhancing.

Local LLMs (and related transformer-based models) can work, they just need a narrow focus. Unfortunately they’re not getting much love because cloud chatbots can generate a lot of incoherent bullshit really quickly and that’s a party trick that’s got all the CEOs creaming their pants at the ungrounded fantasy of being just another trillion dollars away from AGI.

Quite the contrary! They effectively sell their cards at a ~20% discount to a bunch of AI companies by “investing” in the companies for a promise to use that money to buy their cards.

It’s as dumb as it sounds and textbook unsustainable economic bubble behavior, but NVidia don’t care because more sales = more stonks = more money to “invest” = more sales = more stonks = more yachts for Jensen. So what if it makes 1929 look like a walk in the park, it’s not their problem.

On maybe the third day of my first programming job, a colleague pulled me aside and said “don’t give me ‘shoulds’ and ‘probablys’. You need to sound confident so I can know to trust what you’re saying”.

That guy was a bit of a dickhead in general but there’s a lot of truth there. To the question “what’s the expected impact of this change?”, “None.” is a good answer. “Well it should work…” is not useful feedback and a good Operations Manager will rightfully reject the change.

Of course it is better to be hesitant than falsely confident, but far too many (software) engineers hide behind indecisive language to dodge the necessary hard work of validating their hunches. If you didn’t test your shit fully, just say so. If you’re right, say it. Personal ego doesn’t belong in an engineering discussion.

I think the premium thing is a channel option. Some channels consistently have it, some don’t.

Regular YouTube 1080p is bad and feels like 720p. The encoding on videos with “Premium 1080p” is catastrophic. It’s significantly worse than decently encoded 480p. Creators will put a lot of time and effort in their lighting and camera gear, then the compression artifacting makes the video feel like watching a porn bootleg on a shady site. I guess there must be a strong financial incentive to nuke their video quality this way.

I can’t prove it but I think there is a real sociolinguistic phenomenon where Americans are unusually obsessed with “safety through temperature”. Like they hear “fire = no germs” as children and somehow internalize that steaks should be somewhere between well done and burnt beyond recognition, and dishwashers should boil your plates like you’re sterilizing a hospital gown that’s been thrown up on by an Ebola patient.

Soap, bitches. It works. Even at 40 °C (with modern detergent and washing cycles). Good thing too because I don’t want to know how y’all are having sex if you think boiling water is the only decent cleaning procedure for putting things in your mouth.

Technology Connections and Hank Green have been shouting this for a while, but that whole issue is way overblown. Some first gen EVs around 2010 had issues, but every major manufacturer since then has way exceeded expectations on battery lifetime thanks to advanced BMS and thermal controls. Car batteries don’t just rapidly degrade out of the blue, the tech has nothing in common with what’s in your phone. But public sentiment has not caught up because most people think Li-Ion = smartphone = dead after 2-5 years, so second hand EVs are way undervalued. Which is great for buyers.

It’s not like you can’t easily total a second-hand ICE by mechanical failure. Just ask anyone who own(ed) a puretech engine. If you went by manufacturer recommendations, the fucking thing might just eat your timing belt one day and grenade itself. And there’s no way a full engine swap on a 5-10 year old economy car is economically viable.

There’s always something that could go wrong when you buy a car. Unless you get comprehensive insurance and warranty, you need to accept the fact that losing the entire car to an accident, catastrophic mechanical failure, or theft is always a risk. If that’s too much anxiety to deal with, get a lease.

If we’re going to be pedantic, let’s do it correctly.

Even with the blinds shut, a space heater will emit a surprisingly large amount of radio waves (mine actually disrupts USB devices with a small EMP when it turns on, and anyone with an RTL-SDR can tell you those 50 Hz harmonics are rough). Some of those radio waves will penetrate the walls/blinds and a tiny fraction might escape the atmosphere and head off into space. From there some will find their way to interstellar space and potentially drift “forever” (well, until the heat death of the universe or whichever theory you subscribe to; I think at that point saying “the photon never got converted into heat energy” is a good enough approximation).

Speed limits are trickier than structural safety margins because of several factors:

• In some areas, particularly remote areas, the process isn’t very well defined. Sometimes the speed limit will be set by one guy who just felt like that was fine. Doesn’t even have to be an engineer really.
• Standards evolve over time (trending towards lower speed limits) but speed limits only change when a tragedy or major road renovation happens. Where I live there’s sometimes a 40 km/h spread on posted speed limits for similar roads depending on whether they were rebuilt last year or 50 years ago.
• Car culture means drivers hold a ton of political power. There are a myriad of traffic devices that cannot be built not because of practical or financial constraints, only because they would “inconvenience drivers”. Lower speed limits are often one of those. People complain so the government backs down despite engineering recommendations.
• A driver is always liable if they drive too fast for the conditions, not the traffic engineer. That goes to the previous point, with zero penalty for not sticking to the sensible engineering choice, political pressure easily wins out. Hard to argue against a work order when the person signing off on it cannot be sued for negligence.

The upshot is speed limits in my local experience have a lot more to do with the municipality/region’s political climate than engineering standards and safety factors. Sometimes I feel like I could safely go 2x, sometimes the limit is 90 km/h on a two-way one lane road with 30 m of visibility where 30 km/h feels like I’m pushing it.

Oh, fun! The debate over the culinary vs botanical meaning of fruit intersecting with the debate of culinary vs topological meaning of soup.

Breakfast cereal is soup[topological] but not soup[culinary]. It is therefore not a contradiction for it to be fruit[culinary].

Water is debatable, everything else why not. If a recipe is generic enough to call for “vegetables”, you wouldn’t be wrong to include any of those things.

Assuming you like eating chicken, when is it wrong to pair chicken with vegetables? I made a vegetable-mushroom-chicken soup last week and it was delish. Whether chicken is or isn’t a vegetable is an academic concern, not a culinary one.

Try putting mushrooms or chicken in the sangria however and you’ll be rightfully prosecuted for crimes against humanity.

If it goes in soup, it’s a vegetable. If it goes in Sangria, it’s a fruit.

Next question please.

GIMP apologists are what Adobe users think FOSS apologists are like. The principled refusal to even consider fundamental UX principles is astounding.

THANK YOU.

I migrated services from LXC to kubernetes. One of these services has been exhibiting concerning memory footprint issues. Everyone immediately went “REEEEEEEE KUBERNETES BAD EVERYTHING WAS FINE BEFORE WHAT IS ALL THIS ABSTRACTION >:(((((”.

I just spent three months doing optimization work. For memory/resource leaks in that old C codebase. Kubernetes didn’t have fuck-all to do with any of those (which is obvious to literally anyone who has any clue how containerization works under the hood). The codebase just had very old-fashioned manual memory management leaks as well as a weird interaction between jemalloc and RHEL’s default kernel settings.

The only reason I spent all that time optimizing and we aren’t just throwing more RAM at the problem? Due to incredible levels of incompetence business-side I’ll spare you the details of, our 30 day growth predictions have error bars so many orders of magnitude wide that we are stuck in a stupid loop of “won’t order hardware we probably won’t need but if we do get a best-case user influx the lead time on new hardware is too long to get you the RAM we need”. Basically the virtual price of RAM is super high because the suits keep pinky-promising that we’ll get a bunch of users soon but are also constantly wrong about that.

??? Of course you do. Investors don’t just buy their way into hypothetical future profits, they buy control over the company. The specifics depend, whether it’s voting shares or the looming threat of debt collection, but the courts will 100 % enforce investors’ right to demand things from companies.

Furthermore the idea that publicly traded companies have some kind of obligation to make as much money as quickly as possible is a reddit-born myth. Shareholders will bring in a CEO, who will be tasked to do whatever and can be fired from the shareholders at any time. Grievous mismanagement and intentional damage can expose a CEO to legal action, just like intentionally destroying tools can expose a worker to legal action. But a CEO acting in good faith has no other obligation than to fulfill the tasks asked of them by shareholders. The problem is that goes wrong when large shareholders plan to sell their shares and need the numbers to look a little better to sell a little higher. But this phenomenon absolutely happens with PE as well – in fact it’s arguably way worse because publicly traded companies at least have legal obligations of financial transparency. Private shareholders can do whatever the fuck they want, including secretly selling their shares to Evil Inc. for them to strip the company for parts and not a single employee has the right to even know who the majority shareholder even is, nervermind what their plan is.

Easier job to do when you’re actually getting reports.

• Reporting = this breaks the rules please moderate
• Blocking = Fuck them, even if they rechnicly abide by the rules I don’t want them near me
• Muting = I don’t want to see what this person does but don’t want to hurt them beyond that

That’s treading dangerously close to a specific kind of conspiracy theory there.

The EBU has pretty strong financial ties to Israel, not least of which their main sponsor Moroccan Oil. But at the end of the day you can count on the ghouls in charge to always find the coward’s way out and follow the money. It’s the exact same pattern of behavior they had with Russia in 2022, which they initially allowed to compete before backpedalling.

Pretending that this is an issue solely because of some conspiracy is not only baseless, but it also unfairly lets most European broadcasters off the hook for refusing to uphold the set the same hard lines with Israel that they set with Russia. Israel competed because European broadcasters were happy enough to include Israel. Simple as.

It’s very funny that you can get ChaptGPT to spell out the word (making each letter an individual token) and still be wrong.

Of course it makes complete sense when you know how LLMs work, but this demo does a very concise job of short-circuiting the cognitive bias that talking machine == thinking machine.