Yes I tweaked it! The number and copy were mismatched and are no longer!
That worst day is likely an overlapping incidents accounting issue; I tried to account for overlapping incidents in another view but probably failed to port it over there.
Hey! isgithubcooked.com is my site; I made this site in February '26, I think
The contribution graph as an outage calendar idea is a commonly recurring one :). I definitely saw it somewhere else as a static asset first before I made this site.
Also, isn't it a major plot point that replicants are not allowed on Earth?
So Deckard being a replicant operating on Earth is illegal; making his status as a replicant "obvious", by endowing him with superhuman traits (or even self-knowledge) would be ill-advised!
It's challenging to disentangle the effects from AI from the cyclical supply and demand cycle of junior engineers in a post-ZIRP world.
People who are currently junior engineers (i.e. graduating class of 2024+) chose their majors in a ZIRP economy (2021 and earlier) that could not hire enough engineers fast enough, but graduated into an economy that was post-ZIRP and is still actively trimming software engineers from the work force (2024 and beyond).
LLMs only became a factor in hiring in the summer of 2025. But by then there was already a massive oversupply (relative to demand) of fresh CS graduates.
My personal opinion is that the market is not (yet) rough because of AI, it's actually mostly just rough because we're not in ZIRP anymore. AI is maybe making it worse on the margin, and is a convenient excuse, insofar as AI is both easier to scapegoat and reason about.
> When interest rates are high, companies and investors don't take out loans
And yet companies borrowing for AI capex like never before - total amount of debt is staggering. And it’s not only likes of OpenAI/Anthropic, companies which had positive cash flow for many years started to borrow too.
Building multi-decade service life infrastructure on the coast in this era of the Anthropocene actually requires you (from an engineering point of view, not merely a regulatory one) to have a buffer zone of wetlands, marshes and sea walls to absorb storm surges.
You can still choose to read that cynically, but I think the cynical angle is taking moral ecological credit for a pragmatic engineering reality :P
To be clear, I'm not trying to argue that this project is ecologically beneficial or neutral (I think that part is beyond debate for many reasons including the noise pollution you've identified)
I was just trying to say that the marshland restoration stuff is likely not lip service, insofar as they actually do need wetland buffer zone to keep their launchpads above sea level.
Really cool to see the difference in token usage across different harnesses.
These kind of independently-verified data points will be super important for companies (and individuals!) trying to measure RoI on token spend as they recover from letting people go wild on token usage in the past year.
IDK, 30GB is a lot of data when we're talking about text!
Moby Dick, uncompressed, is ~1MB. Compressed, it's around 500KB.
I feel fairly certain that one could fit all of the textual knowledge required to cultivate a world-class in <60,000 Moby Dicks. (Arguably in <5,000 Moby Dicks with intense effort/pruning).
I think a specialized model could squeeze all you need to know about a certain profession in 30GB. But not all professions at once, which is what these models try to do.
Or maybe not, maybe there's a world model needed for human level at any profession that is very hard to quantify and requires more than 30GB by itself.
Yea to be clear I think >70% of the information is not profession specific.
I just think about all the content I’ve consumed in my life to become a professional software developer and I would be very surprised if it couldn’t be adequately represented by <30GB of uncompressed text. Most of the work was in “training”, not really in data.
The “foundational overlap” of K-12 education is identical for most professions with 2-8 years of “finishing” on top.
My mental model is that the budget is spread across maybe 20% genetics (most of our instinctive/genetic information is surely pretty useless), 50% k-12 education, 30% for professionally-specific knowledge.
The human genome is 800MB, (and 500MB with compression according to GPT) so that part is easy.
I think the problem is that the conversation drifted from "all useful human knowledge" to "enough knowledge to do my job".
Both a human and a current large model will bring up unprompted some tax regulation that applies in your district that could matter to your task. A small model won't know the tax code of every city in the world, as it's probably impossible to fit, and it's the kind of thing that unless you already know about, it's very hard to search for, let alone come up with unless specifically prompted about something tax related.
Unless we start having different small models trained for a certain profession for a certain city, I think we'll need those trillion parameters.
> A small model won't know the tax code of every city in the world, as it's probably impossible to fit, and it's the kind of thing that unless you already know about, it's very hard to search for
What makes it hard to search for? If you tell me you want to open a lemonade stand in East Bumbling Falls, I won't know that city’s municipal ordinances. But I will know that cities tend to have ordinances. So I’d search broadly for what codes that city has. Then I’d go through each one looking for anything that might affect a lemonade stand.
Usually at least here it's tens of thousands of laws like "replace article 3 paragraph 2 with:" and it's not like you can ctrl f "lemonade stand" and get something useful.
A strong llm without that knowledge can probably spend millions of tokens parsing every single ordinance and give you an answer if you directly ask about that.
But both a local human or an "all knowing" llm when you just tell it about the stand will be like "hmm wasn't there a law about low calorie beverages being tax exempt? Let me check" and after a very precise search come back with a tip that if you sell sugar free lemonade that's tax exempt.
I mean would you get an accountant from another country just because he can use search?
just curious; which fable/opus versions are you using?
Fable-1m-max-thinking output, despite being incredibly slow and expensive, feels like it bucked a trend towards superficial loquaciousness in their models that had been building since 4.6.
I'm a bit of a luddite when it comes to upgrading models, fable was the first one to make me give up Opus 4.6-1M-max.
Many of the most valuable companies in the world (who, incidentally, are spending the most money on frontier model inference), already have "the expensive part", the labeled dataset.
The cost of maintaining the datasets and the models is getting commoditized by startups like braintrust and huggingface.
