That this feels like such a representative statement is disheartening. Remember, in a functioning free-market (behold. bow.) negative externalities, like socializing the burden of reduced air quality resulting from for-profit enterprises, are considered inefficient (boo. hiss.) and deleterious. Even people with jobs have to breath air.
Of course there is a cost to everyone and the cost is dirtier air. Articles like this don't mention or strongly deemphasize the fact that everything is a tradeoff and a recession or stagnation of growth has horrific consequences as well. No one sees the jobs and welfare that could have been created if environmental regulations were less onerous.
light penetrates in useful intensity for photosynthesis to about 200m so that seems like a decent cutoff for "deep sea" for the purposes of discussion around the unique communities there
Except that much/most of the life in that top 200m regularly commutes into the depths in pursuit of food/shelter/temperature. We all like sea turtles. Leatherbacks regularly dive to well over 1000m.
The article is presented as propaganda and the media format is trash but it raises good points. Still it fails to point out growing use of insect, plant, and byproduct feeds. It also fails to mention that feed conversion ratios in farmed fish are better than any of the land animals commonly farmed, so overall i/o is about as sustainable as it gets for animal husbandry. We also fail to recognize that forage fisheries have short generation times and are easier to measure, so maintaining healthy stocks is easier than trying to do the same with wild caught pelagic fish. Sustainable fisheries are beautiful resources. When managed and informed properly they can afford more production than would otherwise happen under maximum exploitation. They require science based management and compliance of course. But there are a lot of success stories as we have adopted fisheries science and I find it annoying that these articles send people to despondency rather than to grounding their consumption choices in science.
The research project they linked in the article [1] talks about 700 sardines needed for 1 farmed salmon which is not a great i/o but supposedly there's a shift away from fishmeal towards soy, algae, insect protein, etc to where fishmeal is more like 15% of farmed fish diet. Note these are wild fish as input so the notion of farming distinct & preserving wild stocks isn't accurate. Per the comments, there were plenty of NYT readers who were not aware of the downsides so I'm not as critical about it as you. Besides yield there's other considerations like disease spread and genetic degradation to wild stocks, sealice, pollution spillover, and so on.
I hear you. But 700 anchovies per farmed salmon is not meaningful without considering the number of sardines per wild salmon. Eating higher in the food web costs one way or the other. It is all about relative impacts of producing food people want to eat.
> How does the value of LLM-produced 100% branch coverage compare to no-LLM 100% branch coverage?
This is such a salient question. Sometimes (definitely not always) the test suites produced by LLMs are so trivial it's scary. Coverage can be an illusion for sure.
I have been seeing such trivial shit getting submitted and deleting them from the test suites, but it's disheartening to see that the devs didn't have the better taste to see those LLM "tests" are so trivial they border on noop before submitting them...
What new or existing metric would better assess test coverage than line coverage and branch coverage?
I wrote a tool called tert - I guess it's called an agent harness now - to run various test runners and log test output and coverage output to disk. FWIU stripping spaces from JSON does save tokens. It seems like feeding coverage lines-missing maps into the prompt results in better output, better LLM-authored tests.
"Refactor these tests for maintainability and coverage. Use fixtures, mocks, and parametrization"
You can write tests that cover every line/branch without actually verifying the code does what it’s supposed to. Relying on (unreviewed?) LLM written tests that provide coverage may not be a great idea.
We try to use types as much as possible. And if we can't, then lint rules, etc. Basically, get everything as far left in the SDLC as we can.
And, if/where we need tests, we write the source so they are few, high value, and complementary. Like actual unit tests, not complex with stuff like mocks just to generate trivial coverage.
reply