The things you allow the LLM to read will obviously be sent as part of a prompt. You can control that though. Reads are tool calls and you can configure permissions for that or be asked every time the agent wants to read something.
This is straight up just uploading your whole working directory. Not as a LLM prompt, but to a Google Storage.
More or less, but most harnesses will rely on tools such as grep to only read portions of files. For even a small sized repo, only a small portion would be tokenized and uploaded
>For even a small sized repo, only a small portion would be tokenized and uploaded
Only on a per-chat basis. Over time, it'll eventually grab the entire repo, or enough of the "secret sauce" that the rest can be reconstructed with AI.
The countries that do pay for it don’t have significantly higher fertility rates.
I’d say there’s an elephant in the room: Childbirth sucks. If you want women to willingly subject themselves to that, you need either a culture that virtually requires it-
And I want to take a moment to emphasise that I don’t like nor want this solution, and would fight anyone who tries for it.
-or you need to pay them well above the actual economic cost of rearing a child, because the process itself is strongly negative. Yes, having a child itself can be great. Eventually, several years in, when they start to become a person.
That’s true, but if you ask any of the women I know, they’ll tell you they’re perfectly happy to keep it at one.
That’s in Norway, by the way, so not one of those countries where you get zero support.
There are no countries that come close to actually paying for it. There are countries that pay for 5% of it, and in those, indeed they don't have significantly higher fertility rates.
please point out the country which pays for adequate housing and other necessities for a family with 4 children, adequate childcare support (or pays one parent to stay at home fulltime), and sufficient guaranteed sick leave.
What I find interesting is that, it really seems that the opposite is true.
My great-grandparents (and their entire ancestry that I can trace, all the way back to 1807) were punishingly, desperately: poor.
Yet, it seems they averaged a heck of a lot more children than me, or my contemporaries. And their children largely lived to be adults with only a few minor exceptions.
The adults themselves didn't seem to live long though, most records of marriages I have are for 17-18 year olds who were already orphans.
This has always been true. We have talked about the costs of raising children forever.
What is different now, is twofold:
1. Bigger financial impact of having a child, both through less government support, and because more women are working. This combined means that to have a child, often, one of the parents needs to stop working, which severely impacts SoL.
2. Less social impact of not having a child. It is far more common to not have children than it used to be, and so it becomes much more of a choice as to whether to make that SoL sacrifice or not.
I think it also becomes a sort of social contagion - in that you are much less likely to have a kid if you don't know anyone that recently had a kid, it feels like a much bigger leap into the unknown.
Yes. Societies have been using their women's children as a positive externality for generations, but now the logic of capitalism and societal freedom has caught up, and now societies would rather collapse than support women and families economically for the work of childrearing.
I think Derek Thompson is a hack. Based on his media appearances from the Abundance tour, he seems to be very libertarian but in a (neo-)liberal space that pays his bills.
AI has replaced video games for me. And there are plenty of cheaper models that "do it" for me, I don't have to spend $$$$ just for entertainment. I will step up to the frontier for serious work. But if I'm just playing, I'm going for the free stuff on openrouter.
Also, ai art is fine. It looks better than me using paint. That said, there are plenty of foss art pieces and public domain that you can leverage if all you really need is placeholders, and that is much cheaper.
When I’m hiring, a human recruiter (or the hiring manager) reads most resumes.
For us, there is some sorting by basic keyword analysis and we start near the top, but there is no proverbial black box that rejects candidates outright.
If candidates are ignored by humans, it’s not because AI rejected them, it’s because we are starting with candidates earlier in the list and might not make it to applicant 537.
Rather unlikely to be the case, supported by the original article itself here, since if your statement was to be the case they would find that the human generated resume is 100% less likely to be shortlisted.
Obviously it’s not 100% of all human resumes are going to be filtered out, but it’s quite damning that human resumes are more likely to be filtered out just because they didn’t LLM-ify it.
Ukraine has concretely demonstrated their ability to perform deep missile/drone strikes (on the order of ~1500km) onto Russian strategic and economic assets.
Ukrainian drone capabilities in the near battlefield (up to ~20-30km deep) are also not contested. Russian milbloggers will openly talk about the difficulty of massing and movement in that area due to the saturation of drone coverage (and btw, this challenge is more or less symmetric).
So the article is not likely exaggerating any of their capabilities. However, it is exaggerating via omission.
In terms of deep strikes, the question is what the success rate of these missions are, what cadence can they sustain, what's the constellation of Russian lapses that have to line up for a successful strike, etc.
Another known area of weakness (that the Ukrianians are working hard on) is the middle range. How to strike quickly at targets of opportunity in the 50-500km range field. This was/is a capability that things like GMLRS and ATACMS provided, but I imagine the Ukraine is forced to ration those munitions carefully.
From my understanding it seems fairly accurate. A lot of the stuff is pretty public - if drones blow up Russia's stuff they obviously know about it. Also drones get shot down so both sides can have a look at the tech.
The reporting seems a bit misleading suggesting drones kill 30k Russians a month. I think that's total Russian casualties of which drones are a good chunk but not all.
