mattbee

19 ms·

181.
by mattbee 3y ago
I think if you automate paraphrasing from an original work to use that original work on bulk somehow, yes. How do you even automate paraphrasing without training it on lots of original work? It's infringement all the way down.
182.
by mattbee 3y ago
Would automated paraphrasing not be a derivative work of the original?
183.
by mattbee 3y ago
I don't know what a fair settlement would be but I'm looking forward to a copyright-holder suing OpenAI to obtain one. These companies have no value if copyright can be enforced on their training data.
184.
by mattbee 3y ago
Prey is _the_ high watermark for this kind of game, I was so surprised to hear what a commercial failure it was, but it's such a work of art. The sense of exploration, dread & mystery, the weirdness of finding the first ending, the
185.
The Ostrich Algorithm(en.wikipedia.org)
1 points by mattbee 3y ago|0 comments
186.
by mattbee 3y ago
I've used it for personal pics for years. I wish the UI were faster but the self-hosting of everything & automatic storage management is excellent.
187.
by mattbee 3y ago
I mean I've not used it but it's definitely there and ready.
188.
by mattbee 3y ago
This place is ours until about 11.30, we should make more of that.
189.
by mattbee 3y ago
These are both regular phrases in my house.
190.
by mattbee 3y ago
Here's a great parody of Jam by Adam and Joe. I enjoyed Jam but when it misfired this was exactly how ridiculous it came across - https://m.youtube.com/watch?v=0t0Ocau-CUg
191.
by mattbee 3y ago
This DVD extra turned into his hands-down best sketch (and probably the best intro to puzzled Americans) - Peter O'Hanrara-hanraran, live on the scene at 9/11 : https://www.youtube.com/watch?v=1SPWgodul_E
192.
by mattbee 3y ago
Huh - last I looked (10 years ago) it was an asterisk script and a few wav files. It looks like lots of people have copied them for different setups, so I guess that domain isn't by the original authors. But just find the recordings o
193.
by mattbee 3y ago
You don't need AI for that! "Lenny" is years old and wastes telemarketers time with only about 15? pre-recorded prompts played at random. Excellent results at https://www.lennytroll.com/
194.
by mattbee 3y ago
I like to imagine someone opened the letter, rolled their eyes, opened "BadKeywords.txt", added our company name and binned it.
195.
by mattbee 3y ago
This is especially true for brand adverts where the customer would have reached the site organically but you have to outbid a rival because they're trying to bid on your brand keywords. It's a "Google tax" for having a
196.
by mattbee 3y ago
Joel Spolsky's 20 year old post on designing an office is still relevant here. He imagines a lifestyle and tries to create a perfect space and location to enable that: The office should be a hang out: a pleasant place to spend time. I
197.
by mattbee 3y ago
Oh you never lost data with btrfs, but you might give up following the trail of mailing list posts and wikis one fine spring morning when it failed to mount. ISTR hours-long repair operations and a level of desperation that a lot of peo
198.
by mattbee 3y ago
Yup this confused me the other day - tried to search for an article I'd written for Dr. Dobbs Journal in 2001 on a very specific topic. No amount of keywords including my name, technical keywords etc. could find it. I assumed that mean
199.
by mattbee 3y ago
Why not offer to sponsor them and cut all the zeroes? When I built a hosting business that was most of our marketing and it worked really well.
200.
by mattbee 3y ago
LTO5 tapes store 1600GB. Glancing at ebay, drives seem to be $200, tapes are $30. (+the server with a SAS card in ofc). It's always going to depend on your scale.
201.
by mattbee 3y ago
Ahhh I loved this game so much. There's such a lovely human touch to it. I tend to hate puzzle games but this makes you feel like you just found a good fudge each time you succeed.
202.
by mattbee 3y ago
The docs were brilliant for v1, it wouldn't surprise me if they were the spec for a great user experience and the code came second. Despite v2 supporting a very similar config file, the documentation doesn't emphasise that and tri
203.
by mattbee 3y ago
Came here with the same quote in mind :) This whole thing has such "Don't Create The Torment Nexus" vibes.
204.
by mattbee 4y ago
Lena is by qntm, but was featured on Charles' blog. It's very good.
205.
by mattbee 4y ago
No, but I believe a large language model is a work that is 99.9% derivative of its inputs, with all that implies for authorship and copyright. Right now it's just a heist.
206.
by mattbee 4y ago
Good luck to them. AI models are automated plagiarism, top to bottom. None of us gave OpenAI permission to derive their model from our writing, surely billions of dollars worth, but they took it anyway. Copyright hasn't caught up so
207.
by mattbee 4y ago
Your example exposes an edge of Go that speaks to this essay. Yes io.Copy works with any Reader and Writer but dig into the documentation and there's a runtime check for the revised ReaderFrom and WriterTo interfaces that you might a
208.
by mattbee 4y ago
It's good press but they hardly have a business model - I assumed they would have a platform or API to access uniform edge computing stuff. I asked about our local pool which has just had a boiler failure and is raising money to avoid
209.
by mattbee 4y ago
There must be plenty who wouldn't mind being shown the door, and have other options. So if that was a new issue with working for Amazon, and a potential deal-breaker for them, it would be worth having a go.
210.
by mattbee 4y ago
Ooo that's really cool and I'd like an excuse to try it. I had a career in duct tape and prayers so very comfortable in that domain :)

More ›