Most of the times, the bottleneck is mainly the company’s legacy. People are using AI on an old, rigid codebase and are still following processes and team rituals dating back from before the IA.
Too bad the author didn't check Fastmail. It checks all the boxes and on top of that its, imho, the best finished product for mail / cal / contacts and by far.
It may even be my favorite tech product at all. Should I be able to pay for more storage space for the drive (options are very limited), it would approach the perfection.
> There's an unfathomable number of games that were released on the Commodore 64, Amiga, Atari ST, and most famously DOS that were made by one or two people, self-published, and have now completely fallen into obscurity to a point where it's not 100% clear who owns the copyright and who to give money even if you wanted to, which has led to the "probably-illegal-but-not-really-enforced" term of "Abandonware".
Mostly unrelated but if you are looking for this "effervescence" in modern times, don't hesitate to take a look at Pico8 games and Ich.io.
On Pico 8, everything is free (except the one time pico 8 licence).
On Itch.io I bought a ridiculous 400 games pack for a few dollars a few years back.
You have to dig a little but I found more gems in both those libraries than I expected.
Archive.org is full of copyrighted software without known authors. I know it because i'v found a lot of Pocket PC & Palm software from studios/companies that don't even exist anymore.
You are living in another world. I don’t know a single people who is not using ChatGPT, Gemini or Claude. Even my non tech parents are using it everyday.
From what I’ve learned about Jev I feel it’s just a very successful marketing campaign to developers not fully understanding data science (and deep learning). It’s nothing new, been around since 2022? Being local is an extreme advantage lol.
Fast and accurate general purpose classifiers DID NOT EXIST before Jev. You could either use an LLM to get a slow and accurate general purpose classifier, or you could use a smaller model to get a fast and inaccurate general purpose classifier, or you could fine tune your own model that would be both fast and accurate, but it wouldn't be general purpose.
> Apple’s always on watch has a declared 15 second buffer for live audio, and a worse scenario of summarizing your conversations for later
Even if you trust them, maybe as an user you can be ok with that. As a non-user who will talk with people wearing Apple Watches, I disagree being recorded and my conversations with the watch owner summarized.
The reasoning and the result document were done after less than 1 or 2 seconds.
Have Ollama suddenly bought GPU capacity?
reply