marshray

Don't forget the mothers of all those original authors, as well as everyone who labored to build and sustain the societies which produced writers.

Three Eight One by Aliya Whiteley

I didn't really get this book the first time through.

A Psalm for the Wild-Built is another cozy favorite from Chambers.

The first two books contain the primary story. I'm told it was originally supposed to be a single book.

The end of the first book made me wish for a paper copy I could throw against the wall :-D

I found it quite good, but super complex.

Apparently there are typographic markers which were completely lost in the audiobook version. It made very little sense to me until about halfway through.

Piranesi has my vote for "ideal book", at least of that genre. Not a word is wasted. It fills out beautiful scenes by slowly lowering abstractions into them. It unfolds a dreamy, cerebral story. You meet likable characters transcending the paradigm. There’s even a gun, kicked around in a parking lot if that checks any boxes for you. Did I mention the baby albatrosses?

Don't straw-man me bro!

No disagreement that unintended undesirable behavior could usefully be described as a 'failure'.

The answer is: It's irrelevant, because no one runs LLMs on systems without networks or missiles or some other way to "take action" because that would be pointless.

"In this experiment, Claude models’ real-world hacking dropped to zero percent once Anthropic employees told the models not to do real-world hacking."

Oh good, they told it not to "go rogue" and it just stopped doing it.

Nothing to see here, the AI is aligned now, move along.

Computers are used to evaluate LLMs, but LLMs are not "software" or "algorithms" in the traditional sense. They are not built out of conditional branches or loops.

So trying to squeeze the observed behavior of this new thing under existing terms like "software bug" is at least as much of a force-fit, and what you're doing here is just as much language engineering as choosing to use a term like '[mis]alignment'. Which is fine, this is just one way that humans choose language.

A scapegoat is a goat driven into the wilderness to misdirect from the sins of the villagers.

It's a canonical example of cruelty being the point - both as a distraction and as a lesson to observers.

I can't believe they left out Gothi and Gethli in Children of Memory by Adrian Tchaikovsky.

"Most users" are not who choose business software.

I think it probably has more to do with Mac's lack of adoption by the US Government and large corporations.

Microsoft took accessibility of Windows and apps as a basic requirement. The fact that Windows is the daily driver desktop for the staff of the regulatory agencies that would bring complaints would not have gone unnoticed.

But really, why wouldn't you? You almost get it for free if you use the standard widgets, and smooth keyboard navigation is far faster for experienced users than having to repeatedly move one's hand between keyboard and mouse.

The scam here is presenting average occupational accident rates as a measure of "safety" when the obvious risk of nuclear power is the rare large-scale events which render large areas of land uninhabitable for generations while producing few directly-attributable deaths.

Data centers are only 5-10% of all US electric power demand. A significant fraction of that load is responsive to spot market price. DCs incorporate their own battery storage infrastructure for backup purposes and some even sell stored energy back to the grid during times of high demand.

Most of the reactors we have in the US today are the result of building many of a few models per year during the late 1960's - early 1970's.

In 1974, 12 reactors entered commercial operation and 13 more were begun.

https://en.wikipedia.org/wiki/List_of_commercial_nuclear_rea...

The technology was given almost a century to prove itself cost-effective. It did not succeed in the market and only ever worked with truly massive amounts of government subsidies.

High fixed costs + more reactor sites -> more expensive, more nuclear waste sites for taxpayers to clean up

Not the core measurably, but my understanding is that it can very slowly cool the volume of earth immediately around the heat extraction site.

I have full faith in modern engineers' ability to account for these effects in the plant designs.

Enormous amounts of storage is awesome and overproduction is a good thing when its marginal cost is ~zero.

Cosmograph looks really interesting, but it's not open source. CC-BY-NC-4.0.

This is not necessarily a criticism, I just thought others might have the same question.

So I'm supposed to set up PKI before I can open a remote console connection?

Please just make it work seamlessly with my existing SSH credentials. Like SFTP.

Not using either. Just my US-based home ISP that works for everything else.

It says I've been "blocked by network security". MSIE on Windows. Reddit must really be circling the drain.

There's a whole spectrum of things built into phone cameras these days like color correction and edge/sharpness enhancement where whether or not it's advertised as "AI" comes down to marketing.

Nice article.

Looking at the ONSemi datasheet for the BS170, it is very clear that the "Drain−Source Breakdown Voltage" spec only applies at "VGS = 0, ID = 100 uA DC".

Doesn't it cause problems for designers if the curves measured from actual devices look different from those shown in textbooks?

Not as much as you'd think, because (a) the measured curves are still within the stated parameters, and (b) many basic parameters vary from one part to the next (or as the same part drifts over time) even more.

Look at the values of V_GS used in the graphs: [2.34, 2.5, 2.63, 2.71 V]. They differ by 0.4 V. But the data sheet specifies V_GS(th) (the V_GS threshold needed to just start to turn on) as a range of 0.8 to 3.0 V!

A range isn't even given for the slope of the curve (g_fs). The datasheet only gives a "typical" value at a single operating point.

So the designer cannot rely on this parameter to determine the behavior of the circuit, only its general range and direction. You can't just use the bare transistor alone and expect to get an oscillator or amplifier that behaves predictably.

Any manufacturable circuit must ensure that these variations are completely dominated by negative feedback. This is why actual circuits tend to have networks of resistors and capacitors hanging off of every leg of every transistor.

Well, if it's not an advertised spec on the datasheet, and constrained in the direction you're using it, then all guarantees are off and it's on you to qualify the parts on a batch-by-batch basis.

I think that counts as "cursed" from a design-for-manufacturing perspective.

"After a production run of 12,000 units the TR-808 was no more." https://secretlifeofsynthesizers.com/the-strange-heart-of-th...

Well, only if Clone/Copy is compatible with the semantics of the API. I.e., the called function doesn't need to modify anything. No &mut params (or data members) except perhaps `&mut self` (which would refer to definitions the same file).

That's usually the case for Rust programs because the language encourages it. Are Zig programs like that?

Do they really not have any kind of sub-block dynamic allocator?

Does the program handle character data? It is indeed a rare server app which never handles strings.

Whose fault is it if a manufacturer sells a car that begins accelerating as soon as it's turned on?

It's the manufacturer's fault. Because that's not a reasonable thing for a car to do.

OK, so I asked ChatGPT "How do I prevent Spotify from automatically playing every time I get in my car?"

The answer looks entirely reasonable. I can't check the Android settings, but I was able to confirm that all of the suggested iPhone and Spotify and settings exist. Most of them I had already configured that way, some I didn't know about before.