I'd be interested, but they don't even list any anthropic model on their code review benchmark, so I feel like they haven't really tested their benchmark on SOTA models.
Why would I ever connect my fridge to the internet? I cannot fathom any feature on a fridge that would incline me towards giving it the wifi credentials.
Well, I think that justice has been served. The feds' prosecution of Ulbricht was the epitome of throwing the book at someone to make an example, when the government's case was pretty flawed, in my opinion. 10 years is enough time to pay the debt of running the silk road.
I am glad that Ulbricht has been pardoned and I feel like a small iota of justice has been returned to the world with this action.
I have been surprised by the strength of Devin at the beginning of my team's use of it, but lately it has been a letdown. I don't know that we are ready to cancel yet but Devin faces some growing headwind for me, at least.
Wow, this made me really emotional. And even though I definitely did not expect a chill Bali DJ set as the motivational link, it also resonated with me in some way.
I can't think of a more honorable way to move through life. I liken the act of closing shop at this point for Hindenburg to the legend I know of Cincinnatus, the Roman emperor who did the job of emperoring and went back to his fields when it was done.
It also moves me how the team is described, as being from whatever background, but all moved by the same fire. I wish that I could be a part of something like that. What the hell am I doing with my life?
Perhaps I am underthinking things here, but this seems like a good use case to just use UDP here. Save all the power consumed in the original connection set up and just the router will handle NAT re-allocation whenever you phone home, right? Responses to those messages should be properly routed to your device. Building on top of TCP seems like a lot of effort for this network situation
Wow, I have been looking for a resource like this for years. I had put the question more in the form of, what year did they call it in Rome what we now call the year 1 AD? Now I know it's Augustus 30 or something like that.
I would like to see this in a timeline format where all the different eras can be shown along side each other.
The existence of a configuration that limits attestation to a probabilistic phenomenon seems like a very thin foundation to stand on here - if it can be changed to requiring 100% attestation rate in the future I think it will be changed as soon as it is feasible to do so.
I haven't reviewed the proposal enough to see how they implemented that, and if it was done in a cryptographic way that prevents changing to 100%, then that could work. But the fact remains that control of our browsing computing environment is diminishing under this proposal.
I have vouched for this article because it blew my mind.
I had no idea that there are left-leaning, feminist-identifying women who held such views.
I wish this article laid out a more substantive foundation for why the author holds their views. I tried to look into the book she mentioned, and I do worry that there's a bit of a echo-chamber effect going on here, because somehow the book was featured by Tucker Carlson? [1]
But it seems that there's a lot of passion here about protecting women-only spaces. I can respect that.
How can it be possible that there is @AI tagging in groups within the E2EE context? The AI generation must run on server side, probably through an OpenAI API, so the text of the conversation must be available to the backend to be able to generate the AI response. What am I missing?
I have personally taken a tour of the NIF at Livermore. The guide was an old hand, who constantly remarked about the efforts of NIF towards "stockpile stewardship," ie the maintenance of the US arsenal of nuclear weapons. It seemed like NIF was all about the stockpile stewardship first, and fusion research was a secondary consideration.
The capability of the NIF to get positive energy from the energy that they impart on the Hohlraum itself is neat, but I constantly discount any milestones that Livermore/NIF report, because the inertial confinement approach has such higher barriers to commercialization than tokamak style approaches, that I just consign it to "boondoggle" in my head.
Yeah, the lasers could be 20x more efficient, and yeah, they probably could figure out how to pump 10s of targets into the chamber per second, but the energy extraction is just completely missing from the considerations. The engineering challenges are a whole 'nother level for NIF, a big barrier to usability.
I have been using fastmail as a single personal account for the past year or so, so I don't have much to add to your technical assessment. However, I have been considering changing providers because of the political position of the Australian government. I've done a cursory search just now and can't find anything to back up my opinion, but my impression is that they have a pretty permissive stance on gov't access to data or just crappy Internet privacy laws, which I disdain. I have been thinking about making the jump to protonmail as my provider instead.
I find this really cool, kudos to the author for trying to build up an understanding from the actual data.
I find it interesting to see that the median temps in these select cities don't seem to be wandering up that much. I wonder if there is some statistical tool to analyze the slight upward angle on the linear regressions of the medians.
I am surprised that these data do not show a more extreme departure. I believe that we are getting hotter temperatures, etc etc. But perhaps my observations are colored by the media on this - if I were a lizard, would I see climate change?
I think that it has some really nice advantages, but it has some glaring issues that make its use case somewhat limited.
When you are defining some component and you can define exactly the query that the component will need in the same place, that is great.
However, doing normal things like file uploads introduce clunky extra steps like sending a mutation to get back a url that you then make the upload request to. Clunky.
Further, in the past my team has experimented with trying to do multi-model transactions with I think Apollo graph ql library, and it was not readily possible. That was back in 2020, maybe things have improved since then?
I think the first piece of real advice I can give you is to seek therapy. Find someone who listens to you and try to explain what you are feeling to them. These problems are not unique to you, but your version of them is unique and giving prescriptive advice here will likely miss the kernel of your truth.
Beyond that, the most simple advice is the best that I think you can hope for here. Make sure that you are taking care of yourself, deeply. What do you need right now? To eat, to sleep, drink water, go for a run? Listen to those needs and act on them before anything else.
Good luck. You only have one life. Make it the best you can.
I am with you on this. Lately I have noticed that I've googled for an issue, find a low quality site with relevant results, and later discover that it's just a copy of the GitHub issues page from the original project. Why didn't the GitHub issue link make it to the first page of Google and this crappy knockoff, with no link back to the source material beat it? So frustrating.
After the CSAM scanning debacle came out this year, it pushed me towards being far less reliant on any cloud servers and I set up a home NAS, transferred my photos to it, deleted everything off of my iCloud, and discontinued my subscription to iCloud for higher storage, going down to the free 5gb tier.
Ever since then, I've been getting notifications on the settings app about how my iCloud storage is almost out of space because somehow my iPhone backup takes nearly 4.8gb, even though I have disabled the vast majority of app backup permissions.
If you dismiss the almost out of storage notification and the accompanying ad for subscription to iCloud, it comes back a month later.
I feel this author's pain.
Another fishy thing that Apple does is that when I bought this phone, they gave a free year of Apple TV+. Most of my paid subscriptions I keep a close eye on and cancel right after I sign up, so that I won't have to remember at the end of the subscription period. It works fine with any app on the app store, I get the service for the desired period. With Apple TV +, if you cancel your subscription to avoid autorenew, it immediately cancels your access to the service.