The product works pretty well and already feels more natural to use than the web UI of ChatGPT.
This landing page is a bit barebones for my taste though. I would've loved to know about what data gets locally processed, privacy assurance, and what gets sent to OpenAI. A lot of bad things can be implied from the two three sentences like: "Chat about email, screenshots, files, and anything on your screen."
That criticism is warranted but I feel that having a "goal" of books to read is a very useful tool to develop the habit of reading in people that never read.
As for the type of people that do not typically read, I would encourage them to read what they would love reading (even if its NYT Best sellers) until they love reading.
I don't have the source, but I remember reading on Reddit that someone had his free programming YouTube tutorials series uploaded to udemy and behind a paywall.
My experience is with the other end, I have been querying Uber drivers about their feelings about a possible upcoming entry of Lyft in my province's market.
So far, on my small sample 10<n<15, it has been unanimous that they would adopt Lyft since they have a reputation among drivers of taking a smaller margin. Some of the drivers also told me that they plan to serve as drivers for both at the same time.
From a technology point of view, one of the collateral benefits of space exploration is that the technology and science that will come from that effort can be used elsewhere.
The problem is that, according to your documentation, local music has to be on a desktop (the only way to listen to local music on mobile is to be connected to the same WiFi of a desktop that serves that music (!)). Which I think defeats the purpose of having your music on the go.
One of the only reasonable defenses I know of (assuming that your picture is already stored in the attacker's db) is this https://cvdazzle.com/.
Even so, it doesn't scale well, given that you do not have the guarantee to fool more than a given set of algorithms (the technique above targets specifically OpenCV haarcascade face detection algorithm).
Still, they heavily limited the scope of actions of both the player and the agent. It's not like the research space consisted of all the items/actions available in Dota 2.
Can you elaborate on how does it improve on traditional OCR?
Other than that, pretty neat that it keeps (or even improves) on the existing formatting on the note.
[1] - https://web.archive.org/web/20251008204636/https://www.rolli...