My favorite Costco story is when it opened up in my town, we joined. We were there one day and bumped into some very good friends of the family there (since passed sadly).
They were talking about how they were admiring all the services offered to members, and said they considered buying a cruise vacation package, but then immediately realized “we’d just be on the ship with people from Costco”
We have a cherry blossom tree. It bloomed a week earlier than last year. We’re not in Kyoto but I did notice and it’s a bit strange. I also noticed some other blossoming trees that typically bloom for about a week, went green after 3 days.
Could it be the incoming "strongest of all time" [1] el nino weather pattern that is expected this year with high probability ? Usually it is only the eastern pacific is mostly severely impacted, but maybe there's a correlation
And human civilization entirely sprung up during it, all of our nations, our cities, our pastures, our lives are built on the ice age. We need to start cooling the world down and we're doing the opposite
The population explosion of the 20th century, and our modern way of life in general, has been powered almost entirely with fossil fuels. The lives of everybody living today and all our societies including 3rd world ones is built on coal and petroleum. Except the North Sentinel Island people.
It will not be possible to keep things going they are, no matter what actions or inactions we need or should take.
People have been claiming for at least a quarter of a century that coal was dead, killed by much cheaper renewables. Today, we will consider ourselves very lucky if we have only just (i.e., in 2024-25) passed global peak coal. All those people and "experts", wrong. Repeatedly and totally wrong. Why? There must have been terribly bad data, bad models, bad economics, or bad assumptions they had been using. I have heard very little in the way of acknowledgement about that or any effort to find and fix the root causes of such failures. Are today's claims still coming from these same flawed approaches?
And it's not just coal of course. Coal was the proverbial canary in the proverbial coal mine because it was supposed to have been killed off long ago. But there's gas and petroleum and we are a decade from global peak carbon even by presumably the same kind of wildly optimistic / flawed projections.
And just passing peak carbon is not the goal. 1990 level carbon emissions were considered catastrophic and we're nearly double that now. Getting it down to well below those levels is just so uncertain and such a long way out that nobody really knows what that will take or how long it will be.
Making solar panels in Chinese coal powered factories doesn't just magically fix everything. Just like it didn't 25 years ago when they said it had killed coal.
You appear to have invented a totally different history to the one we just lived through so you can be grumpy about it.
Why was coal going to be phased out ages ago?
I remember a global scramble by some of the best minds on the planet to drive the cost of solar below the cost of coal because that was the only way they saw coal being displaced.
Like Google's Renewable Energy Cheaper Than Coal, or RE<C. Launched in 2007, or the US government's DoE Sunshot initiative in 2011 to reduce the cost of solar by 75%
It is in fact the mindless coal is dead parrots who have been living in their own worlds for the past 25 years. And they're the ones who appear to be grumpy when that's pointed out.
Solar is in actual fact covering the whole of increased electricity demand globally, as of 2026, and still growing fast.
It covered 75% in 2025. And wind the rest, also still growing.
We could have got there faster, but there was, and is, a long running disinformation campaign to deny there was a problem, question the root causes of the problem and demonise the most promising solutions.
So that just happened now, this year? Solar is only barely covering just growth? Decades after it had "killed coal"?
So it took solar that long to take the easiest 8% of global electricity generation share. The long-dead coal is somehow still far a larger producer of electricity. 4x bigger than total solar capacity.
"Coal-fired power generation in India in April increased to 164.9 average gigawatts, compared with 160.7 average gigawatts last year, according to data shared by S&P Global Energy. According to the data, coal-fired power generation rose sequentially by 5.6 average gigawatts, or 3.5%, in April."
Longer periods can be called paleoclimate. As you may have noticed, most types of humans did not exist in previous climates, and we are unfamiliar with the conditions of those time periods, much less if we were to bring them upon ourselves in a period of time that isn't even capable of being shown on the chart you've chosen to use.
I'm just clarifying parent comment that "1200 years of data is climate" by saying that longer periods also are climate data. I could have posted a graph of the holocene as well (I don't know that it would materially change my point).
I made two points. The other was that we are in an ice age.
Normally, discussions of climate refer to the last 12k year interglacial period as having come out of an "ice age". You're using the broader geologic term referring to any presence of any polar ice cap as an "ice age", which would cover the last 3 Million years. So what you're saying is that in the 300k years homo sapiens have never existed outside of an "ice age" and that the our speciation (eg in savannahs of Africa) was driven by the many glaciations of this current Ice Age? Even homo habilis hasn't been around that long.
That's saying that since the continents and earth's currents haven't changed, we're in the same age, AMOC is a minor technicality, and the oceans would need to rise to the straight of Panama to be significant.
No, climate is based on consistent weather data over a long period. Across long enough periods the underlying assumptions that make climate a meaningful thing to talk about fail due to orbital mechanics etc.
Plate tectonics for example shows you can’t even assume an area’s latitude is consistent, just look at the fossil history of Antarctica. Humans have dumped so much carbon and methane in the atmosphere even 100 years ago was quite different.
It very much reads like you feel like you need to offer those particular points here to try to diminish concerns about global warming informed by the 1200 year Kyoto cherry blossom record. Is that not the case?
Yes I am diminishing the case of global warming about the tree. The tree they kept records of 1200 years ago is not the cultivated one seen in many parks now. That tree is about a million years old. Article just says "Kyoto’s cherry trees", which would include the old and new ones. Importantly, the new cultivated ones bloom earlier.
This could be a case of trying to make a climate argument, when the underlying data is more nuanced. Maybe we should just say it's nice the trees are blooming earlier.
- The climate has *always* changed. It’s been warmer. And yes, it’s been cooler. There is nothing abnormal about the climate changing.
- There is actually very little scientific proof that the current up tick, is human-made. Yes, there’s correlation with the Industrial Revolution, but that’s all it is atm, correlation. There’s little verifiable proof. It’s speculative. It’s a theory. Yes, there’s overwhelming consensus, but that’s still doesn’t make it fact. And consensus has been off target plenty of times in the past.
- “The science” isn’t always as fact / truth based as it would like us to believe. Scientists are human too. Egos, career aspirations, groupthink, jealousy, etc. The scientific method is a stunning standard. Unfortunately, it’s implemented / executed by humans, flawed humans.
There’s three sources exemplify #3, of course there are others.
While the climate has always changed and there's nothing abnormal about that, it has never, ever changed anywhere near so radically in such a short period of time; the rate is what's abnormal. XKCD has a fantastic visualization of this:
So pair that with the correlation with the Industrial Revolution/increasing carbon dioxide levels in the atmosphere, and with the verifiable scientific fact that carbon dioxide works to trap heat...and surely you can at least see why there's overwhelming consensus, right? What would compel you to operate as though this isn't the most likely explanation for the unprecedented rate of warming we're seeing?
We better hope we're the cause of the warming, because then we conversely have a shot at slowing it or stopping it. If we are incapable of causing a change of this magnitude, then the actions we are taking to slow the change would likely be ineffective too, in which cause coming generations are in for a world of hurt.
As such, it always strikes me as bizarre when people question human contribution to climate change without by extension freaking out far more about the urgency of taking drastic action.
>> then the actions we are taking to slow the change would likely be ineffective too.
Many countries are taking steps which are mitigated by many developing countries who rely on cheap energy to grow and build out of their third world status.
So yes, on the one hand a lot of countries are doing something but will it ever be enough to counter other countries continuing to pollute at unprecedented levels? I don't know.
I have, and the dramatic change correlating with the rise of human release of co2 makes it highly improbable that it is natural. But as I said: If it somehow is natural we're far more screwed so we better hope it isn't, because if it is natural the cost and resource impact to protect us against the effects would be far more brutal.
But…you're correlating previous climate changes (which had much slower rates) with "Even though this one measures daramatic'ly different, it's the same.".
As some others have asked you, would you be so kind as to please suggest other sources for either the required energy inputs, or the required reduction in heat losses, so as to provide other plausable explanations for the available data?
We agree that corelation is not causation. I suppose we should also agree that ignoring a correlation when choosing what to investigate would not be science.
In this very thread, two posts up, the direct parent of the comment you're replying to, I gave you a link to a visualization of the climate record and asked you to look at it, and pointed out that the sudden and unprecedented rapid rate of change since the Industrial Revolution is precisely what leads us to believe this isn't natural.
You responded by insisting (without evidence) that "the climate record over the looooong term simply is not that accurate". And now here you are telling someone else to "look at the climate record", the climate record that shows precisely what they're saying, the very same climate record you cast dispersions on moments ago, hoping to somehow trick them into not believing their own lying eyes. You're not operating in good faith.
I’ve looked at it. I’ve seen it before. I saw it when it was initially published. It’s why I said: the planet has been warmer, and it’s been cooler. There is no normal.
So when I hear “climate scientists” make big brush statements about the last 100 yrs, 500 yr, or even 1000 yrs, I laugh. When they make a proclamation about “since records were kept” I laugh again.
I laugh because I’ve seen that graph and I know they evidently haven’t seen it, or are for some reason pretending otherwise. I presume they’re trying to be funny. They’re certainly not being scientific serious.
- Climate models are demonstrably incorrect, but the current consensus is that it is changing faster than expected
- if the climate models are demonstrably incorrect, how much weight should we really put into it for understanding our impact on the model?
- if cherry blossoms blooming earlier are an input into the climate and this is a surprise that is not modeled, how much are we missing in our model? do you know how insane it is to try and model secondary/tertiary effects. Climate change is not an independent variable, it is in fact also recursive, therefore any input we are missing (and we are demonstrably missing many) fans out in our model and makes it inaccurate in either direction
- take for example what was being discussed in 2023 https://www.sciencedirect.com/science/article/pii/S0304380023001680
- as a comparison to what we think we know now https://www.nature.com/articles/s41558-026-02598-w
- if we can't even get this right, what do you think our chances of understanding the implications of the impact to everything indirectly dependent on phytoplankton?
- For the sake of discussion lets follow the premise that you align with: Climate change is human made, we should do what we can to prevent or slow down climate change.
- to what end? everything we do affects the climate. we aren't just changing the climate, we *are* the climate. there have been many species that were so effective that they meaningfully changed the environment and climate and eliminated other species or themselves along the way. should we eliminate them too? do we actually understand what we are doing if we do that?
the presumption that we are changing the climate in a way that makes it unsuitable for humans has not been demonstrated thoroughly. we can see that because population is still increasing, but therein lies the paradox.
if you're arguing humans should stop human made climate change then we should go further and just say that there needs to be fewer humans. if you're correct, human made climate change is the mechanism for getting to fewer humans.
if the model is demonstrably wrong and if it were trivial to correct and understand what was wrong about it then we'd have an accurate model by now, but we don't. after all your premise that ~100 years of observation is enough to say anything, how come we can't get the model right? If you follow this far out enough it also says we have no idea how to reverse the trend. if we did our model would not still be wrong in 2026.
the climate and life on earth is changing, will always change, there will be winners and losers, while we try to change it so that we are the winners, we have more evidence that it's an inherent property of the system and that life will continue on despite change than evidence that pre-industrial climate was the ideal climate that we should strive towards. If you want to argue that we should strive towards an environment where humans have the most success, well the evidence points to the fact that we are currently in that climate. population is increasing. just know that there are many losers in the climate you wish to promote. also know that ANY climate has a lot of losers. that much has been clear.
It's almost certainly caused by man as all of the evidence suggests that it is. But if it's not, that's a much more serious problem since if it's some unknown natural phenomena we probably can't do anything to stop or slow it from happening and we don't know how hot it will get or how quickly. Maybe humans have triggered a yet to be discovered tipping point and there's no stopping it.
Some reports are already saying that global warming is progressing faster than predicted... maybe we're on an exponential slope to higher temperatures and don't know it yet.
Simple enough… share links to the science that thoroughly proves what you’re saying. “Al Gore said it, it must be true,” isn’t going to cut it.
As for the “never ever”, that’s another assumption. The climate record over the looooong term simply is not that accurate.
Of course, there is a fair amount of correlation and circumstantial evidence, but parroting that as absolute fact and causation does make it scientific. There is a lot of “telephone” on this issue. Those using hundreds of years or even thousands should be met with skepticism.
To clarify: I agree 100% that the climate is changing. It always have. Unfortunately, the rabbit hole of proof is not that deep. Most of “proof” is based on consensus and groupthink dictated by “the experts.”
You mean those who have dedicated their careers to learning the science and studying the various evidence and using that as a base of reference?
Who would you cite otherwise?
Those providing the alternate view to "the experts" all seem to be seriously lacking in any related qualifications. Politicians, radio personalities, fossil fuel industry lobbyists, marketers, advertising executives. They're all pundits, they're not experts, they literally don't know what they're talking about.
If you believe everything, or even most things, those various groups tell you, well, I'm not sure how to help.
The most important Science [1] in the history of Science and history of The World… is no science at all. It’s consensus. Nothing more.
Let that ^^^ sink in.
I might consent, as a matter of caution. But I do so reluctantly and with skepticism. Most importantly, I’m not fooling myself and thinking I’m believing in something that is. Not. There. That is, Science. I’m not mindlessly parroting a narrative [2] - like a bad version of kindergarten telephone - because I’m afraid of asking obvious, worthy, and intelligent questions. I’m not a victim of my biases.
You said all you said and yet didn’t site anything. How is that? But this is exactly the problem. People take a position, bias kicks in. They double down. More bias. There’s no turning back.
Thank you. You’ve proven my point.
Why do you think “the experts” are any different? They’re likely worse, as they have more to lose.
> To clarify: I agree 100% that the climate is changing. It always have [sic].
Did anything I said imply that I didn't recognize that? There's no need for you to clarify, I was very clear myself that I was responding to your lack of understanding about the nature of the changing climate.
The XKCD comic I already linked shows you the climate record for the past 22000 years; you can see with your own eyes that the rate of change since we began burning hydrocarbons on an industrial scale, particularly since the year 2000, has been precipitous and looks like nothing that's happened in the entire history of humanity as depicted. Maybe you missed it - it's a little subtle - but the comic already lists its sources on the side. At a high level, they determine the climate record by examining things like tree rings and ice cores; if you're curious, those sources are happy to explain their methodology in detail. Beyond that, do you have a specific reason for casting doubt on those sources?
So let's see - going back to my comment, I pointed out the climate appears to be warming up faster than ever before; I (well, XKCD) has given you sources for that. You yourself acknowledged that there was a correlation between the warming climate and the Industrial Revolution, but I suppose we need a source for the correlated rise in carbon dioxide. Here's a graph from climate.gov using data from NOAA, ETHZ, Our World In Data, and the Global Carbon Project. If you visit each of those sources (which are linked to from the graph) you can then drill down and how they themselves synthesized it (as we know, that's how science works).
So at this point we have evidence of a phenomenon (the planet suddenly warming much faster than before, per the XKCD visualization and its sources) and we have a demonstrable mechanism (the warming effect of carbon dioxide, per Eunice Foote's experiments circa the mid-19th century) matching our data (the increase in carbon dioxide, per NOAA and ETHZ and Our World In Data and the Global Carbon Project) that solidly explains it. That looks like science to me. Further skepticism without any contradictory evidence and you're just getting into poor epistemology frankly, and I'll just have to start throwing around metascientific ideas like Occam's razor and Russell's teapot and post-critical logic. You keep demanding more "science", but what does that actually look like to you? You look at the entire scientific community (who you scare quote as "the experts") and their body of work and mindlessly dismiss it for not being "scientific" enough. It's a rhetorical feint, not genuine intellectual curiosity.
To be honest with you, I haven't even seen An Inconvenient Truth myself. Have you? Does Al Gore just come on-screen and assert things, or does he give explanations for you to ignore? Sources for you to ignore?
If my tone is short, it's because it's both frustrating and amusing to be treated with such airs of intellectual condescension when at this point the evidence and consensus for anthropogenic climate change is so strong. When everyone who's devoted their life to understanding something says you're wrong - even if you think you may actually be right! - it'd probably be better to argue from a position of humility, because the odds are very good that you are in fact wrong. As Carl Sagan said: they laughed at Columbus, they laughed at Fulton, they laughed at the Wright brothers. But they also laughed at Bozo the Clown.
A few years ago I was walking through Queen's Park in Toronto and saw a Flat Earther accost people with the same sort of arrogance. Most laypeople walking through the park on a sunny weekend afternoon, it turns out, couldn't tell you off the top of their head about how we know that the Earth is round; whereas Flat Earthers performing a stunt are more than prepared to tell you why they know that the Earth really is flat. And I guess there's something valuable about court jesters making people aware of how much common sense they take for granted, but that doesn't make them any less worthy of ridicule.
In fact the repetition of consensus is a classic tool of propaganda. “I’ve heard it a bunch of times… it must be true.” And if you don’t think scientists aren’t subject to the same influence, you’re naive.
The argument that tries to belittle the current warming by pointing that in the past there have been times when it was warmer, does not have any value.
Besides the fact that a much lower temperature than in the past can do a lot of damage to the present-day beings that are not adapted to such temperatures, those higher temperatures from the past had been stable for long periods.
Now, when such a higher temperature from the past will be reached, there is no guarantee that the warming will stop there.
The greatest danger of the current warming is that, if we do not do anything about it, nobody can currently predict with any certainty that it would stop at a still acceptable temperature, matching a temperature from the distant past, instead of continuing towards even higher temperatures.
As a human, I do tend to mostly care about the period of the Earth's history that has allowed humans to exist. I'm sure the Mesozoic was nice, but I wouldn't want to live there!
I actually live in a city that usually gets over 50 degrees Celsius for weeks and we pay a LOT on energy just to keep the A/C on. I don't think it is sustainable for the whole world and I'm afraid of what will happen here once we reach higher temps.
The linked picture is just an artist's interpretation, not a photo, and I assume you were mostly joking too. I think we agree that it's not a good idea to hurry in warming the planet.
A better time range would be the average species lifespan of the plants and animals we eat. Too short a range highlights noise; too long a range highlights unrelated data.
Weird that there are any downvotes. The current administration in the US, spent its first few months literally pulling down and replacing data on US government websites. These included many categories of scientific data, including most especially climate data.
HN was literally in an uproar, some users were trying to back up data. And yet now we see people posing such questionable data from US government websites?
Please. No output from the US should be trusted at this point. It's all ideologically tainted.
Perhaps they take issue with "_any_ other source". I agree that U.S. govt data is now suspect, but there are far sources. Maybe an s/any other/more reputable/ would be accepted?
> A 1,200 year time series.. that's definitely in the climate area.
Superficially, sure, but this is not a good dataset for any climate related argument. Cherry trees live about 100 years under optimal conditions, so you’re talking about multiple generations of trees here, with significant adaptation and selection along the way (humans heavily influenced the development of these trees, and the current “standard” tree for cherry blossoms in Japan is actually a hybrid first created in the 1700s). In short, even if you set aside measurement reliability and consistency over time, this dataset is heavily confounded.
As an aside, you’ll note that the primary change is that the lagging tail of the distribution is pulling forward (i.e. the distribution is getting narrower) not that the distribution overall is shifting forward. You can find trees blooming “this early” many hundreds of years ago, just not as often as now.
But wouldn’t that be true for all periods in the dataset? You see ups and downs over the centuries, and in each of those centuries I’m sure humans heavily influenced their evolution. Then you see a pronounced upward trend that just happens to also coincide with what we know to be serious, sustained and highly unusual planetary-wide warming.
Sure, but it doesn’t matter. Confounding means that there’s an uncontrollable factor in the analysis. You can’t just assume that the uncontrollable factor is the same over all measurements in the dataset.
Just to illustrate the point, the trend you’re describing (which you haven’t correctly described, but I digress) could be due to warming, or it could be due to selection in favor of earlier blooming varieties in the last generation. You can’t know which explanation is true, so you can’t draw conclusions.
We say the same thing about southern California. When the forecast is the same for 350+ days out of the year, that's not weather, that's climate.
I say that as someone from Texas that lived in LA for several years. Texas weather changes by the hour and this time of year it is advisable to keep an eye on it. In LA, you could go weeks without checking the "weather".
That does not make any sense to me. SoCal is famous for having stable, comfortable weather (and hence its high land price). Some places have volatile weather, some places have more consistent weather.
This is incorrect, and is a very common misunderstanding of what the term "anecdote" means and what the actual problem with anecdotal data is.
The dichotomy is between "anecdotal evidence" and "scientific evidence," and the important distinction is not that the latter simply has more data points than the former. The critical distinction is about the methodology used to gather the data, not merely the number of data points gathered.
Climate is also dimensional. Kyoto is a point. A point over time is a line, a line through a 3d set of data. That a single point is seeing an effect is interesting but not as significant as widespread changes. Only when multiple measurements create a 2d map of realtime data, which becomes a 3d bulk over time, should we draw conclusions. Sadly, that is also happening. But the later should be the topic of conversation, not a single very visible data point.
The single visible data point is interesting, as an illustration.
It doesn't prove climate change one way or the other, but that is a discussion that ceased to be meaningful decades ago. Climate change is real, it is significant, and it is caused by humans. Further arguments about that are a (deliberate) waste of time.
Having accepted that, and dismissed the time-wasters from the conversation, we can look around for things that we notice. One of them is the way it affects the times that trees bloom, giving us an opportunity to discuss the way that affects other aspects of the ecosystem.
That, in turn, helps inform conversation about just how important the consequences are. Unlike the fact of climate change, it's not obvious how much the consequences matter to us, and what should change to avoid them. That is a conversation worth having, but it has been impossible while we're still listening to people reciting decades-old falsehoods.
Interestingly, if you have one-dimensional observations f(t) of a k-dimensional strange attractor, the lagged vector time series [f(t); f(t - tau); f(t - 2 * tau); ...; f(t - (k - 1) * tau)] maps onto the full k-dimensional attractor. Specifically (as I check Wikipedia) it's a diffeomorphism, an isomorphism of differentiable manifolds.
Presumably the earth system isn't at anything resembling an attractor right now, but I wouldn't be surprised if people are trying to use related techniques to try to detect qualitative changes in the system dynamics (like bifurcations).
Maybe someone more knowledgeable could chime in on whether/how measurements at a single point on the earth's surface might be used to do that?
Nonsense bills get introduced all the time. I’m not saying this shouldn’t be taken seriously, but this eventually getting codified is a long shot.
There are so many issues with how this can work in practice. Best case it just asks how old you are like a website that shows mature content, and the user lies. So from a liability perspective that shifts it to the user who gave false information. Beyond that there’s no practical way to actually verify someone’s age at the OS level.
Age verification in general rolled out so fast, I'm more inclined to think that nonsense gets passed more easily if it enables going after drastically more control.
The government doesn't care if it's nonsensical and contradictory and enforcement would be a dumpster fire because all that essentially grants their buddies (if not now then in the future) in the executive more power.
I'm happy I invested in setting up Codex CLI and getting it to work with ollama. For the toughest jobs I can use Github Copilot (free as an academic) or Gemini CLI. If we see the per token price increase 5x or 10x as these companies move to focusing on revenue, local models will be the way to go, so long as stuff like Gemma 4 keeps getting released.
There was a headline saying they were, and the actual article showed they were doing nothingbof the sort.
If you read HN headlines, and don't even bother to click into the comments and see everyone calling out the headline as bogus, you might think something like your statement is true.
I just looked into this a bit because I thought he still had some kind of role at Microsoft even after leaving as CEO/chairman, but it turns out that in 2020 he left any and all positions at Microsoft as it was investigating him over inappropriate sexual relationships he had with Microsoft employees.
Before that he had a role as a technical advisor and sat on the board of directors.
I also found it interesting that Steve Ballmer owns considerably more of Microsoft than Bill Gates (4% for Steve Ballmer while Bill Gates owns less than 1%).
He still visits Microsoft occasionally. A friend showed me a picture of him visiting Microsoft in Beijing a few months ago (he was excited about BillG visiting). So my guess is that he still has an interest in Microsoft products.
I couldn’t know, but generally speaking, older billionaires don’t typically interact with the world in the same way most of us do (well, those without a social media addiction anyway). The device is someone else’s problem.
He’s still around as a part time advisor, he has to officially step back or no one would take Satya seriously, but on important stuff like AI he is a bit more active.
If he doesn't use Windows, you won't hear about it. And if you hear that he uses Windows, it might not be true. He loses nothing by denying it. If it worked for his friendship with Epstein, it will work here.
They made a movie to make money. I doubt anyone holding the purse strings cared one iota if that bit were corrected or not. It’s not really a retcon either because they didn’t change anything.
I disagree that evaluation is always a coding task. Evaluation is scrutiny for the person who wants the thing. It’s subjective. So, unless you’re evaluating something purely objective, such as an algorithm, I don’t see how a self contained, self “improving “ agent accomplishes the subjectivity constraint - as by design you are leaving out the subject.
Sure. There will always be subjective tasks where the person who asks for something needs to give feedback. But even there we could come up with ways to make it easier / faster / better ux. (one example I saw my frontend colleagues do is use a fast model to create 9 versions of a component, in a grid. And they "at a glance" decide which one is "better", and use that going forwards).
OTOH, there's loads you can do for evaluation before a human even sees the artifact. Things like does the site load, does it behave the same, did anything major change on the happy path, etc etc. There's a recent-ish paper where instead of classic "LLM as a judge" they used LLMs to come up with rubrics, and other instances check original prompt + rubrics on a binary scale. Saw improvements in a lot of evaluations.
Then there's "evaluate by having an agent do it" for any documentation tracking. Say you have a project, you implement a feature, and document the changes. Then you can have an agent take that documentation and "try it out". Should give you much faster feedback loops.
> Things like does the site load, does it behave the same, did anything major change on the happy path, etc etc.
I asked Claude to build a web app to run locally polling data from the LAN. It fought me for four rounds of me telling it that the data from the api wasn’t rendered on the page. It created tests with mock data, it validated the api, it tested that the page loaded. It was gaslighting telling me that everything worked every time I told it that it didn’t work. I had to tell it to inspect the dom and take screenshots with Playwright to make it stop effing around. I don’t think it ever would have found the right response on its own.
Even after deliberate intervention, it regressed a few rounds later and stopped caring that tests failed. Whatever, I don’t treat it as anything more than a sometimes-correct random output machine.
In science there are ways to surface subjectivity (cannot be counted) into observable quantized phenomena. Take opinion polls for instance: "approval" of a political figure can mean many things and is subjective, but experts in the field make "approval" into a number through scientific methods. These methods are just an approximation and have many IFs, they're not perfect (and for presidential campaign analysis in particular they've been failing for reasons I won't clarify here), but they're useful nonetheless.
Another thing that get quantized is video preferences to maximize engagement.
They were talking about how they were admiring all the services offered to members, and said they considered buying a cruise vacation package, but then immediately realized “we’d just be on the ship with people from Costco”
reply