There was a startup that did this in the mid 2010s named Magic, but was just via SMS. I used it a few times to get random things done, and it was really useful when it was cheap, then it became mega expensive.
There's two seperate things DoorDash seems to be doing: "Tasks" in the physical world (taking photos of inventory on shelves, closing Waymo doors), and then some seperate app for training AI models.
As for Magic, they were an SMS-based virtual assistant. They still exist today. They went downhill. https://getmagic.com/
Why?… The experiment.yaml shows that it is calling h100/200 explicitly, it’s pretty common for humans to say “number bigger more gooder” for anything… Lie and reverse the values and see what happens. I would put money on a rabbit hole of complaining about it being misconfigured.
It means holding the actual stocks in the underlying index, as opposed to synthetic replication, which aims to achieve returns matching the index via derivatives or other techniques.
It's physical in the sense that literal means not literal nowadays.
ETF and index arb traders use the term physical to describe securities that require full margin. Example: Sell stocks, buy index futures (and reverse) is the classic EFP equity trade. To be clear, futures are highly leveraged, thus do not require full margin.
I have tried most of the major open source models now and they all feel okay, but i’d prefer Sonnet or something any day over them. Not even close in capability for general tasks in my experience.
Harvester is just Kubevirt with some UI atop it, the same as Redhat Virt. Works fine if you’re hosting datacenters or whatever, haven’t seen it be suitable in smaller manufacturing environment
The country is perfectly capable of having its own rotten morals, and outsourcing of all blame to Israel is just excusing the mistakes of American leadership.
> outsourcing of all blame to Israel is just excusing the mistakes of American leadership.
Isn't it really the other way around? Israel is literally outsourcing its war and its war crimes to the US military (the strike on the girls' school was not Israeli but American).
Sure, Israel is getting some bombardment, but the lion's share of retaliatory strikes are being borne by American allies, almost all of whom have now lost trust in the US, and are now being forced to buy stockpiles from the EU and even Ukraine because the Americans came unprepared.
It's telling when the tightly controlled media in those countries lets billionaire magnates openly criticize their country's relationship with the US (not Israel), on national print.
If your friend plans on killing a hooker because he likes the idea of snuff porn, and you pitch in to kill her coworker and her boss so there are no witnesses and so your friend doesn't get hurt and because okay maybe you also enjoy snuff porn when you're in the right mood, then it's a joint venture and you share culpability. Trying to divy up fault (45% or 55%?) is kind of besides the point. Trying to decide who's ultimately responsible (0% or 100%?) is both besides the point and violates every ethical principle we have.
Trump sent half of the US's fleet to the Persian Gulf to mount a war on Iran, in part to distract us from the Epstein Files, in part because he thinks he's a czar who we'll title "The Great" for his territorial expansion since the Nobel committee vetoed "The Merciful". Rubio said that we knew Israel was going to attack Iran, and we would have stepped in to defend Israel from the counterattack, so we decided to just attack Iran ourselves. Hegseth ("Deus Vult" Hegseth) and the general staff apparently are disseminating the view internally that "Trump was anointed by Jesus to light the signal fire in Iran to cause Armageddon and mark his return to Earth", and they are His blessed instruments.
In your preferences there is a local apps and hardware, I guess it's a little different because I just open the page of a model and it shows the hardware I've configured and shows me what quants fit.
I haven't seen a page on HF that'll show me "what models will fit", it's always model by model. The shared tool gives a list of a whole bunch of models, their respective scores, and an estimated tok/s, so you can compare and contrast.
I wish it didn't require to run on the machine though. Just let me define my spec on a web page and spit out the results.
reply