I want it to commit on my behalf, and I don't pay full attention before commits. But I do review every line at the PR level, just as I would if some other human were to submit a PR, and if I find something off, then I ask it to fix it before I will approve the PR.
re: automatic coauthoring -- you can turn that off, but I prefer to turn that on. Commit ownership is something that I want to make clear was not just a human, but a human + AI. Agentic coding does get special treatment from me, because the volition to make changes doesn't come from me. I want that to be clear in the history. The only time I do my own commits (without the coauthor line) is when I make some change myself, without a coding agent.
Really interesting article but there's one thing I wanted to point out about the reasons for code review:
> If a company were shipping bugs at, say, a hundredth the rate we were at Centaur while relying primarily on review to catch bugs, then I could see their point, but that's not what's happening at the typical software company where people don't want to move away from human review [there might be non quality (as in non bug rate) related reasons to keep human review, such as keeping a high bar for code quality or keeping the codebase human-understandable, which pretty much immediately stops being the case if you let a fleet of agents go wild on a codebase] because of the perceived risk of shipping bugs.
I don't agree that the main point of a code review is quality assurance; there are other reasons that have more to do with team convergence and junior staff learning: https://www.embeddedrelated.com/showarticle/807.php
Really liked it, and somewhat envious that you were able to create such a well-written, cohesive piece of just about the right length. (I blog on different topics and it's so hard to keep things short and not wandering.)
I've used these while working as a volunteer in a camp kitchen during work weekends, where we had maybe 20-40 guests eating a meal. You put dishes/silverware/pots/pans/whatever in a square tray, push it into one side of the Hobart, and pull down a lever to close the doors and start the cycle. They take a number of minutes that you can count on one hand... I seem to remember 90-120 seconds but I'm not sure. Stuff comes out clean and hot on the other side, and you want to pipeline this so it's one person with dirty hands feeding stuff in, and one person with clean hands (and thick gloves) pulling the square trays out, letting them cool, and putting the clean dishes/etc. away.
They release a huge amount of steam and they're wonderful for this kind of volume (20-40 people at a meal + all the cooking implements to support this). Larger than home use, though --- unless of course you had a mansion with servants. And I get the sense that maintenance and operating costs would be a lot higher than a regular plain dishwasher.
"The geography caveat matters. A data center in the Scottish Highlands and one in Phoenix, Arizona, face very different realities. But even in warmer climates, the shift toward 45-degrees-Celsius coolant moves operators significantly closer to that chiller-less ideal — where chillers may turn on just a few days a year when the outside air temperature demands it."
Chandler is about the same (50-something% SRP, 30-something% Colorado, < 10% groundwater). Some groundwater use is sustainable, by the way; the cities do calculations of how much water seeps into the ground from rain or irrigation, and the idea is you're supposed to pump up less than that amount, minus a 5% "cut to the aquifer" to make sure there's extra margin to avoid depleting the aquifer.
> For the City of Phoenix (the political city, plus about half of Paradise Valley), the tap water breaks down like this: roughly 60 percent originates as Salt and Verde River water, delivered through the Salt River Project. A little under 40 percent originates as Colorado River water. A very small amount is groundwater, pumped from local wells. That is the portfolio. The number that gets cited in most coverage (41 percent groundwater, 36 percent Colorado, the rest split) is the statewide figure, the whole of Arizona averaged together. It is not Phoenix. Phoenix is its own arrangement, and the arrangement is unusual.
p.s. good article, by the way. It highlights a lot of the byzantine quirks of water management in Arizona.
> In an interview, McCalip says his initial rough calculations a few years ago suggested that data centers in space would cost in the range of 7 to 10 times more, per gigawatt of capacity, than their terrestrial counterparts. “It just wasn’t practical,” he says. “Not even close.” But when Elon Musk began publicly backing the idea, McCalip revisited the numbers using publicly available information about Starlink’s and Tesla’s technologies and capabilities.
This is a good start, but the calculation doesn't include office space and overhead (for every 100 developers there is maybe 5-10 support staff to cover the additional legal / administrative, and don't forget the extra cost in supervisor time to manage them)
Largely those datasheets are poorly written... but in practice it doesn't matter much because the manufacturing variations of offset and gain error are larger than the quantization level.
The datasheet specifications are largely standardized across ADC manufacturers and do not represent the exact mathematical norm but rather the range of ADC performances. So it's not that "the transfer characteristic is always mid-tread sampling" but rather that "the ideal model of the transfer characteristic is always X".
I do have some historical examples where mid-riser was used, but in practice it doesn't matter when you consider ADC variations and the variation of gain and offset are multiple counts. (8-bit or more; this exact definition would have been important for ADCs less than 8 bits.)
So what does this mean for those of us who have been using Claude or ChatGPT with agentic coding? Is there anything we can use to reduce costs with Mellum2?
Dammit, I have an 80%-written article covering the same issue but for ADCs, and had to put it aside for the past few months. There's historical precedent here from the 1960s and 1970s, and in large part it involves testing and definitions of gain and offset error in ADCs.
https://twitter.com/jms_embedded