If we use the strict definition of organic results in SERP, these aren't the result of webpage indexation, they're the output of widgets and other natural language parsing in Kagi.
I evaluated gpui for a project I’m working on. While I’ve found the rendering and performance to be great (and actually switched to Zed as my daily driver after playing with it), the (absence of an) accessibility story made me grudgingly just use tauri.
I really want to love gpui and would like to avoid using webviews, the advantage of web-based apps is that accessibility is (relatively) easy to do if you’re attentive to it. At the time I looked at it, gpui did not support screen readers etc at all.
I suppose I understand why they made that decision given their business priorities, but it’s hard for me to tell my vision-impaired friends and colleagues that there’s literally no way for them to use my app.
> Two-thirds [of participants] reported medium to high user satisfaction with the device.
I don’t know much about medical trials but this seems surprisingly low to me, especially given that the study population is presumably predisposed to liking the device (since they opted-in to an experimental study).
Did the implant not work in these cases or were there other quality-of-life issues? I wish university press releases on science were less rah-rah and presented more factual information. I guess that’s what the NEJM article is for.
It seems that it’s black on white, forms, and reading involves adjusting zoom and brightness until you focus a single word at a time?
And it still uses your regular peripheral vision so the experience merging the two might be uncomfortable.
Not discounting the success at all, but anything messing with your senses is probably very hard to adapt to unless it’s pretty much a perfect match with the experience you’re used to.
In college, I worked for a small team in a large organization that used SAP. My team tracked everything in an Excel dashboard, and I was tasked with automating data ingestion from SAP into Excel. The only tool I had available was the SAP GUI input emulation API for VBA. It was extraordinarily painful to set up and would break every time the SAP team would change the GUI to add or remove a button. Lots of fun.
To be fair, it is almost impossible to get data out of SAP. Their "security" is all there to prevent users from doing anything useful. You have no access to the underlying database (like DB/2, oracle), you have to use their GUI or write a custom ABAP program.
But in most cases, the functions you need to call to write data to disk are usually closed off to developers due to "security". If you have access to the database (almost impossible to get), the data for the important tables are spread throughout multiple tables with names that look like names created from /dev/urandom.
I am familiar with SAP, all that you say is only true if you don't know how to do it. I see this a lot, people that are familiar with "normal" technology, try to invent ways to do things in SAP. What is wrong with writing an ABAP program, or using a provided communication/interfacing method to transfer data? If you are stuck on GUI/files/DB level, sure you wont be able to do anything. You can for example generate complete excel files on the server, no need for the GUI. There is a running joke on SAP forums about how many times excel file generation was invented.
Some of the cryptic table names date back to R/2, sure, but they are the de-facto standard data model for those business data. If you have business systems communicating, for example product, business partner or financial data, it will have a mark on it of how those data are handled by SAP.
But then there are CDS views (for some time now) that have long descriptive names, and metadata to help you make database queries. You are not meant to read or write database tables directly (as of some time).
Again, I will say, where I worked, due to "security" lots of items were disabled, even for developers. The only item that you could use to get data was SE17, some people (a few) were allowed to have SE16. But due to memory, getting data that way was very slow.
Also, SE16 did and may still have security issues. That was the reason for it being disabled for 99% of the users.
While it’s interesting that Dom0 avoids Spectre-style branch prediction attacks it’s not clear from TFA exactly why that is so. How does the architecture of the hypervisor avoid an attack that seems to be at the hardware level? From my limited understanding of Spectre and Meltdown, swapping from a monolithic to a microkernel wouldn’t mitigate an attack. The mitigations discussed in the VMscape paper [0] are hardware mitigations in my reading. And I don’t see Xen mentioned anywhere in the paper for that matter.
I guess it’s sort of off topic, but I was enjoying reading this until I got to the “That’s not just elegant — it’s a big deal for security” line that smelled like LLM-generated content.
Maybe that reaction is hypocritical. I like LLMs; I use them every day for coding and writing. I just can’t shake the feeling that I’ve somehow been swindled if the author didn’t care enough to edit out the “obvious” LLM tells.
I think the author actually meant "Yes, vmscape can leak information on Xen, but only leaks from a miniature Dom0 process." Leaking from an small pool not being a security issue they seemed to consider.
Agreed on the point about hw-level mitigation. The leakage still exists. Containing it in a watertight box is quick and effective, and it does avoid extra overhead. But it doesn't patch the hole.
it might be as simple as more rigid context transfers flushing caches. there are a lot of guesses on here now. itd be great if people stopped using may or might and looked in the code. everyone's hopping on the lack of context and adding guesses. thats not helpful
Please see my other comment where I share more details about VMScape and why Xen is not affected. In short, it is because branch predictor state is flushed when transitioning to Dom0. Indeed, it has nothing to do with type of kernel...
And yes, LLMs were at work. The "quote" in the article is not an actual quote...
It's not the em dash, but the negative parallelism ("not X, but Y"). This is a pattern which some LLMs really like using. I've seen some LLM-generated texts which used it in literally every sentence.
(The irony of opening with this pattern is not lost on me.)
As an aside, Wikipedia has a fascinating document identifying common "tells" for LLM-generated content:
I'm also on the spectrum and like using various kinds of parallel construction, including antithesis.
I also tend to use a lot of em dashes. If I posted something I wrote in, say, 2010, I'd likely get a lot of comments about my writing absolutely, 100% being AI-written. I have posted old writing snippets in the past year and gotten this exact reaction.
I originally (two decades ago) started using em dashes, I think, because I also tend to go off on frequent tangents or want to add additional context, and at the beginning of the tangent, I'm not entirely sure how I'll phrase it. So, instead of figuring out the best punctuation at that moment (be that a parenthesis, a comma, or a semicolon for a list), I'll just type an em dash (easy on a Mac).
Then I don't go back and fix it afterward because I have too many thoughts and not enough time to express them. There are popular quotes about exactly this issue.
It's a kind of laziness in the form of my expression to give me more mental capacity to focus on the content. Alt 0151 and Alt 0150 are still burned into my memory from typing em dashes and en dashes so often on Windows.
I suppose I'll have to consider this my own punctuation mode collapse that RLHF is now forcing me to correct.
I've started deliberately using em-dashes and “smart” quotes (made easy by configuring a compose key) — mostly because they look nice, but also out of spite for any software that's somehow not properly Unicode-aware in 20-fucking-25.
Does using Grammarly count as AI-assisted writing?
I use Grammarly because it helps fix speech recognition errors. One of the challenges of speech recognition use is that it is a bit difficult at times to construct grammatically correct sentences in your head, then speak those sentences, and then proofread them before you start the next bit of writing.
When I was starting out in rust, replacing my IDE’s `cargo check` invocation with pedantic clippy (which has a lint for this use of `into_iter` [0]) was very useful in learning these parts of the language.
My grandfather, with whom I was very close, suffered from Parkinson's in his last decade or so. For a long time he was doing OK: Occasional confusion and the slow, shuffling walk that is characteristic of the disease.
One day he had a minor operation that left him needing a wheelchair for what we thought would be just a few weeks. But he never regained his strength and was never to walk again, which led to a steep and sudden decline in his mental condition. It was truly devastating to see one of the sharpest people I knew become an angry and confused simulacrum of the man I so admired.
I wish I had realized two things then: First, as you say, maintaining mobility is the crucial to the well-being of the elderly. Second, immediate physical/occupational therapy after a fall or surgery is essential to people at risk of losing mobility. Sadly it wasn't offered to us and we didn't think to ask.
My dad is going through that shit right now. He fell a few weeks ago and hasn’t walked since.
I live abroad to make more money and feed my ego and I only see him 3–4 times a year. On top of that selfishness, every now and then I catch myself selfishly thinking I don’t want to go through that, which makes me feel like an even worst piece of shit.
Man, be easy on yourself. You already have the world to put you down, no need to add to it. Life is complicated and I’m sure that you trying to have a better life and a career is not just for ego. Love yourself a bit.
reply