zarzavat

17 ms·

421.
by zarzavat 8mo ago
The reason that FPV drones are so easily disrupted is that they are too light to carry anything more than a radio and fly low. Disrupting the signal for a normal-sized aircraft is much harder. If you're flying at 10s of thousands of fe
422.
by zarzavat 8mo ago
You mean it's equally fraud as that. Fraud has a particular definition which does not permit "someone else is committing fraud and getting away with it" as a defense.
423.
by zarzavat 8mo ago
Not everybody is using their headphones on the go. 99% of my headphone use is at my desk while I work. Wired is more convenient than wireless since it's one less lithium battery to charge. It's true that humans are not particularl
424.
by zarzavat 8mo ago
You can give the macro-structure using stubs then ask the LLM to fill in the blanks. The problem is that it doesn't work too well for the meso-structure. Models tend to be quite good at the micro-structure because they've seen a l
425.
by zarzavat 8mo ago
Yeah prompting doesn't work for this problem because the entire point of an LLM is you give it the what and it outputs the how . The more how that you have to condition it with in the prompt, the less profitable the interaction wi
426.
by zarzavat 8mo ago
Consumers don't read spec sheets. My mother doesn't even know the difference between RAM and SSD - it's all just "memory" right? But she knows that when she goes to the Apple Store the computers are built to an impr
427.
by zarzavat 8mo ago
To err is human. Let's embrace our humanity in the face of this proliferation of insipid perfection. I want the raw tokens straight out of your head. Even if they are lower quality, they contain something that LLMs can never generate:
428.
by zarzavat 8mo ago
The manufacturers don't care about display quality, because displays are hard and expensive. Apple has enough volume that they can get a custom panel. Users on the other hand, they definitely care about display quality more than they c
429.
by zarzavat 8mo ago
Arguably with SIP a hardware indicator light is not strictly necessary, the OS could force the indicator pixels to be lit.
430.
by zarzavat 8mo ago
Yes for the basic plan. However there are people who claim to use the API and spend hundreds, or thousands, of dollars a month.
431.
by zarzavat 8mo ago
I'm not getting results. That's the point. Claude doesn't fucking work without human intervention. When left to its own devices it makes bad decisions. It writes bad code. It needs constant supervision to stop it from going o
432.
by zarzavat 8mo ago
There's a real disconnect. I was talking to a junior developer and they were telling me how Claude is so much smarter than them and they feel inferior. I couldn't relate. From my perspective as a senior, Claude is dumb as bricks.
433.
by zarzavat 8mo ago
It feels like everyone's gone mad. Here I am mostly writing code by hand, with some AI assistant help. I have a Claude subscription but only use it occasionally because it can take more time to review and fix the generated code as it w
434.
by zarzavat 8mo ago
Let's just do leap minutes. If humanity survives long enough to witness a leap minute without destroying ourselves then that's ample compensation for the minor inconvenience.
435.
by zarzavat 8mo ago
I don't know about Pichai in particular but whenever rich people do things that don't make sense the answer is usually tax avoidance or financial engineering. If you have many shares concentrated in a particular company that you c
436.
by zarzavat 8mo ago
Filesystems have a property that changes preserve locality. A change made to one branch of the tree doesn't affect other branches (except for links). Databases lack this property: any UPDATE or DELETE can potentially affect any row dep
437.
by zarzavat 8mo ago
The thing about dream logic is that it can be a completely rational series of steps, but there's usually a giant plot hole which you only realise the second you wake up. This definitely matches my experience of talking to AI agents and
438.
by zarzavat 8mo ago
This is not an "AI error". This is a human decision to use known unreliable AI for waging war, in full knowledge that civilian death is an inevitable consequence. If you decide your strike locations using a pair of dice, it&#x27
439.
by zarzavat 8mo ago
It's like having an early/broken chess engine. An amateur with a chess engine that blunders 10% of the time will hardly play much better than if they didn't use it. They might even play worse. Over the course of a game, those
440.
by zarzavat 8mo ago
This is exactly how all other weapons of mass destruction were rationalised. "If we develop <terrible weapon> we can save so many lives of our soldiers". It always ends up being used to murder civilians.
441.
by zarzavat 8mo ago
What are we supposed to talk about in this thread exactly? The developers of this model are evil. Are we supposed to just write dry comments about benchmarks while OpenAI condones their models being deployed for autonomously killing peopl
442.
by zarzavat 8mo ago
I fondly remember the browser in F-script Anywhere was fantastic for debugging your own apps and reverse engineering other people's. Smalltalk is how computing should have been: an open book. I hope that one day we can rediscover that
443.
by zarzavat 8mo ago
Honestly I believe they did it because Tim Sweeney has fuck you money and he got pissed off at Apple.
444.
by zarzavat 8mo ago
If the school is wealthy enough to provide free laptops, then you're right they're going to go for the cheapest option. But if the school expects the parents to provide laptops, then the parents are more likely to choose this.
445.
by zarzavat 8mo ago
I believe we are talking about slightly different things. Yes if they thermally coupled the body to the processor, then a small patch of the body would get very hot, burning the user. However, the fact that the aluminum gets hot during prol
446.
by zarzavat 8mo ago
I swear that AI could independently develop a cure for cancer and people would still say that it's not actually intelligent, just matrix multiplications giving a statistically probable answer! LLMs are at least designed to be intellige
447.
by zarzavat 8mo ago
It's essential for thermals. Without the unibody, it would throttle sooner and you'd lose performance.
448.
by zarzavat 8mo ago
Silent corruption has been a feature of Time Machine for the last 19 years. But haven't you seen the new glass effects, isn't it cool?
449.
by zarzavat 8mo ago
It's not exactly a bump if they raise prices at the same time, though with the RAM situation I'm not mad.
450.
by zarzavat 8mo ago
CSV quoting is dialect dependent. Honestly you should just never use CSV for anything if you can avoid it, it's inferior to TSV (or better yet JSON/JSONL) and has a tendency to appear like it's working but actually be hiding

More ›