It's interesting that the shading for the white pieces is preserved. I wonder if a dark mode extension is affecting the colors.
1,544 karma · joined March 7, 2013
It's interesting that the shading for the white pieces is preserved. I wonder if a dark mode extension is affecting the colors.
"Alive" pieces remain permanent throughout the game. In both 2.16 examples, white can capture black by filling the gaps.
"Eyes" only have one space each and are fully surrounded by a single color. In both 2.16 examples. there are no eyes. Look at 2.17 with this new definition to see where the gaps are not yet eyes.
2.17 (3) is asking you to place a white piece within the black formation. This renders correctly for me in firefox.
Usually if you see an id in your http logs you can simply search your database for that id. The v4 to v7 indirection creates a small inconvenience.
The mismatch may be resolved if this was available as a fully transparent database optimization.
On the podcast, Durkin frames improved detection as a positive, because it means people will get the care they need.
Maureen Durkin, an epidemiologist at the University of Wisconsin-Madison, just presented a paper [1] (currently in peer review) studying 8 year olds from 2000 to 2016, categorizing and counting autism severity over time. The most severe cases were unchanged, or decreased, and the largest change was in those with no measurable functional limitations.
This unpublished paper suggests that identification of children with milder symptoms is the strongest driver.
[0] https://podcasts.apple.com/au/podcast/autism-the-real-reason...
[1] "Trends in the Prevalence of Autism By Adaptive Level between 2000-2016: Evidence from a Population-Based Sample of 8-Year-Old Children in the United States" S. M. Furnier and M. S. Durkin
It's a scaling strategy. The payback period is reduced to 0 if a customer pays for the vehicle. That capital can be fed back into manufacturing, lowering costs, and expanding market penetration.
Collections of SVGs for the icons: https://github.com/markmanx/isopacks/tree/main/collections
Paying for my own ads felt similar to shopping at a local bookstore: I paid extra for the culture I wanted to see. There's a market for it, but, you're right, it probably wasn't big enough to justify its existence at Google.
1. They don't fall out, because they "hook into" the shape of your ear.
2. They sit much closer to the inner ear, so they feel invisible when side-sleeping.
3. It's a perfect fit every time, so you don't have to worry about noise leakage or discomfort caused by pressure.
If you don't want to spend $200 on a pair of plugs from an audiologist (I'd strongly urge you to reconsider!), the best off-the-shelf alternative for me was silicone plugs. To solve the side-sleeping problem with silicone plugs, you can tear them in half or use children's versions depending on your ear canal size.
Not to nitpick, but "better" isn't on a single axis. Taking a photo is a better experience than searching and keying in every component, even if the accuracy is identical.
AI tools don't even have to be "vastly" better. They could be, and often are, even worse on several axes, because they often trade quality for ease-of-use.
You may assign a lower weight to the ease-of-use axis (as engineers, we tend to), but then you're likely not the ideal customer for many of today's AI products.
While it's true that you could avoid signing cookies, this isn't the default for any server library I'm aware of. If your library doesn't require a secret to use for signing, you should report it.
I'm also unaware of JWT libraries that default to "none" for the algorithm (some go against the spec and avoid it entirely), though it's possible to use JWTs insecurely.
His lab has shown functioning eyes on the backs of tadpoles, allowed frog leg regeneration where none existed, and performed several other modifications that change the communication between cells to trigger desired growth. Surprisingly, the interventions are point modifications, then the system handles the rest of the process.
Cell-to-cell communication has a lot of explanatory power for a cell (or collection of cells) "knowing what/where to be".
Apple - alleged Siri eavesdropping: $95M [0]
LinkedIn - alleged unauthorized ai training on private messages: ?? [1]
Google - alleged unlawful data collection in Texas: $1.4B [2]
[0] https://www.usatoday.com/story/tech/2025/05/11/apple-siri-95...[1] https://www.itpro.com/security/privacy/linkedin-faces-lawsui...
[2] https://www.businessinsider.com/google-alphabet-settlement-t...
"I previously opted out of model training by writing to the support team. Will you continue to honor my opt-out?
Yes. If you opted out by contacting support or using our privacy form, your account will represent that request."
https://help.openai.com/en/articles/7730893-data-controls-fa...
In 1996, they optimized an FPGA using a genetic algorithm. It evolved gates disconnected from the circuit, but were required.
The circuit exploited the minuscule magnetic fields from the disconnected gates rather than the logical connections.
Humans are capable of reproducing copyright illegally, but we allow them to train on copyrighted material legally.
Perhaps measures should be taken to prevent illegal reproduction, but if that's impossible, or too onerous, there should be utilitarian considerations.
Then the crux becomes a debate over utility, which often becomes a religious debate.
1. https://github.com/dudykr/stc - Abandoned (https://github.com/swc-project/swc/issues/571#issuecomment-1...)
2. https://github.com/kaleidawave/ezno - In active development. Does not have the goal of 1:1 parity to tsc.
1/3 = 0.333 repeating
3/3 = 0.999 repeating
1 = 0.999 repeating
The conclusion was 0.8 g/kg (or 0.36 g/lb) was enough. 1 g/lb is ~2.2 g/kg
https://blog.google/products/gemini/google-gemini-deep-resea...
At one point all git operations were down: https://news.ycombinator.com/item?id=42691184