modeless

44 ms·

811.
by modeless 1y ago
How certain are you that none of your printers or visitors' laptops or whatever were ever compromised by a botnet? Or that your ISP isn't also serving customers who are compromised or malicious on IPs adjacent to yours? A few hard
812.
by modeless 1y ago
The safest option is a hardware security key because it is not vulnerable to phishing. And I expect Google to trust it above all other forms of authentication because of that. So anyone who is worried about losing access to their account sh
813.
by modeless 1y ago
Yeah I get nagged once every few months, maybe. Easy enough to ignore.
814.
by modeless 1y ago
Yes, a lot of places don't let you remove the phone number. But Google does.
815.
by modeless 1y ago
I would start looking at the networks you are using. You may be unknowingly sharing your public IP or IP block with compromised machines that are part of botnets, which makes Google (rightfully) very suspicious of logins coming from there.
816.
by modeless 1y ago
If you want to prevent SMS from being used, remove the recovery phone number and/or 2-step phone number from your account. That's how I've had my account set up for many years, to prevent SIM swapping attacks. Just make sure
817.
by modeless 1y ago
I hope they don't build agent-specific interfaces. I want my agent to have the same interface I do. And even more importantly, I want to have the same interface my agent does. It would be a bad future if the capabilities of human and a
818.
by modeless 1y ago
Agents respecting robots.txt is clearly going to end soon. Users will be installing browser extensions or full browsers that run the actions on their local computer with the user's own cookie jar, IP address, etc.
819.
by modeless 1y ago
Seems like Firefox will ship WebGPU on Linux before Chrome does, then.
820.
by modeless 1y ago
If I understand correctly, each level of the hierarchy divides its input into chunks of variable size, but outputs a fixed amount for each chunk. The chunking is learned. The model can choose to compress data by making its input chunks bigg
821.
by modeless 1y ago
I don't know if this is the one but something like this is clearly the future IMO. We need more levels of hierarchy to efficiently generalize to longer sequences with high level structure. Back when Byte Latent Transformers came out I
822.
by modeless 1y ago
I can't believe this is such an unpopular opinion. I don't think HN is a place for me anymore.
823.
by modeless 1y ago
If that's true today it was true the day they made the original offer. Microsoft didn't change that.
824.
by modeless 1y ago
This would also happen if OpenAI developed the same thing internally, right? I don't see how not acquiring them improves anything.
825.
by modeless 1y ago
I don't understand why Windsurf would care after they've exited.
826.
by modeless 1y ago
But these specific tools are similar in that they can't cut skin so they are completely safe for kids, and can make curved shapes in cardboard relatively easily (more easily than scissors at least).
827.
by modeless 1y ago
For the price you could get five of these cardboard cutting tool kits. https://www.make.do/products/makedo-discover
828.
by modeless 1y ago
They said they're training a new base model for better multimodal performance soon. I wouldn't expect it to be able to read an image like that today. Maybe if you provided it in text format.
829.
by modeless 1y ago
It's available now
830.
by modeless 1y ago
Seems like it is indeed the new SOTA model, with significantly better scores than o3, Gemini, and Claude in Humanity's Last Exam, GPQA, AIME25, HMMT25, USAMO 2025, LiveCodeBench, and ARC-AGI 1 and 2. Specialized coding model coming &qu
831.
by modeless 1y ago
It's been ten years since the last settlement was finalized in those cases. They were a success and we don't need to wait any longer to decide that.
832.
by modeless 1y ago
Yes, class action lawsuits in general and this one in particular present strong disincentives that companies respond to, especially after they lose.
833.
by modeless 1y ago
You linked to an example of statute preventing bad behavior. Yes, it took years, the law moves slowly, but it happened. Retroactive compensation was provided for those affected, but most importantly, corporate policies changed.
834.
by modeless 1y ago
There's no doubt in my mind that Silicon Valley's success should be credited in part to California's ban on non-competes. Bad for individual companies, but good for industries. Just look at Meta's current poaching spree,
835.
by modeless 1y ago
ARC-AGI 3 remindes me of PuzzleScript games: https://www.puzzlescript.net/Gallery/index.html There are dozens of ready-made, well-designed, and very creative games there. All are tile-based and solved with only arrow k
836.
by modeless 1y ago
How much latency you can perceive greatly depends on the context. But in the right context humans can perceive display latency down close to 1 ms as demonstrated by Microsoft Research many years ago. There is no excuse to be skeptical about
837.
by modeless 1y ago
Fil-C is a modified version of Clang that makes C and C++ memory safe. It supports things you wouldn't expect to work like signal handling or setjmp/longjmp. It can compile real C projects like SQLite and OpenSSL with minimal to n
838.
by modeless 1y ago
If ChatGPT is causing this, then one would expect the rate of people being involuntarily committed to go up. Of course an article like this is totally uninterested in actual data that might answer real questions.
839.
by modeless 1y ago
Not mentioned in the article is that increasing regulations for safety and efficiency are the cause of worse visibility, because they encourage or require more airbags, more strength, more crumple zones, higher hoods, and more shallowly s
840.
by modeless 1y ago
Strongly disagree. Optimizing for resource usage above all else was the wrong thing. The developer experience was beyond awful. Terrible tooling. And poor resource isolation meant that it didn't recover from errors during development,

More ›