6,620 karma · joined March 5, 2012
Twitter: @runeksvendsen
GitHub: https://github.com/runeksvendsen
Consultants are obviously making huge amounts of money implementing LLMs for companies. The question is whether the company profits from it afterwards.
If you already have the answers to verify the LLM output against why not just use those to begin with?
As I understand, it's a lot of work because there's no common way to target these different GPUs. Is this correctly understood?
If I don't know better I'd think you just cherry-picked the prompts with the best-looking results.
You create one on your iPhone and another "backup" key from a desktop PC running some open source software. If your iPhone breaks you can always use the other.
Similar to a server configured to accept multiple different SSH keys.
How can this be possible in a non-centralized manner?
Because the entities that have the power to outlaw them are their biggest customers.
That's one way to put it
Isn't a Dockerfile just a sequence of dependencies, rather than a tree?
Their landing page[1] says
> Rama is programmed entirely with a Java API – no custom languages or DSLs.
But this sure looks like an embedded DSL with stringly typed identifiers.
The 2013 BMW 1 Series had a starting price of $35k. A Tesla model 3 starts at $40k in today’s dollars. Seems pretty even.
I think Apple hasn't done this because it's not practically possible.
For example, if it's possible to change the content of a view (which it obviously should be), then it's possible to remove all content and and replace it with all new content. If triggered by some user action, this effectively is "going to a new screen". But the OS-level gesture won't work since the developer in question didn't use your proposed OS API call "createNewView", but instead just replaced the content of the current view.
However, Apple could make use of this OS API mandatory in order to pass review. Although I still think there's some subjectivity around what is "a new view" (that I can "go back" from) versus "an existing view with stuff changed" (that I can't "go back" from).
Safari is bottom left.
Hmm, so if I understand this right, the query
SELECT Orders.OrderID, Customers.CustomerName, Orders.OrderDate
FROM Orders
INNER JOIN Customers ON Orders.CustomerID=Customers.CustomerID
would return two tables: (1) an Orders table with only the OrderID and OrderDate columns present; and (2) a Customers table with only the CustomerName column present — without any connection between CustomerName and either Orders.OrderID or Orders.OrderDate?EDIT: Or, perhaps both Orders.CustomerID and Customers.CustomerID would be present in the respective tables since they're referenced by the JOIN?
“Real” meaning “adjusted for inflation”.
I would love to be able to say e.g.: this column is either a PK in the “results” table or a PK in the “errors” table.
I wish Postgres had a stable, more low-level API than SQL. Perhaps something that resembles the query plan I get from EXPLAIN. An application written against this API could implement an SQL DB: (1) parse SQL (2) implement a query planner that spits out a query plan written against this API.
IANAL, but I would think it'd be the job of the business in question to hire lawyers, who would then offer their opinion on the matter.
Then an email dump of 45 GB of useful information could instead be 4.5 TB (with 1% useful information), and wading through all the non-information to find something useful will not be worth the time of the adversary. The more important information you have in emails the more you need to increase the misinformation-to-information ratio.
As a client, you do know what to ask for: everything you support, listed in order of preference.
Why would you, as a client, care whether the server supports some format that you don't understand?
No, I don't think so. Why do you think so?
> I feel like a good API is limited in what it accepts, and this alone is enough to say one should not do content negotiation unless forced to.
Content negotiation does not change what the API accepts. It changes the format of the response.
Side note: I guess a better title for this blog post would be "Don't share endpoints between machines and humans". If the consumer of an endpoint is a human, it's probably a bad idea to make machines use the same endpoint. Content negotiation is fine, for example if I want to have an API return binary data instead of JSON (if, for example, I'm on an IoT device with sparse resources). This is fine because it's exactly the same API, with just the data format being different. And in both cases it's consumed by a machine/computer.
In the case of the frontend/HTML, the consumer is a human (and not a machine), which — as the article mentions — adds a bunch of constraints that are not applicable to machines: a need for pagination (because scrolling through a thousand results can be annoying), a need for ordering (because I want the most relevant results first), displaying related content (as the article explains). The machine API doesn't necessarily need any of these features.
I'm not following you. Do you have a better example?