trollbridge

6 ms·

31.
by trollbridge 18d ago
These “breakouts” are simply marketing.
32.
by trollbridge 18d ago
Nope, because I think releasing open weights is far more important and realistic than ever expecting the big American AI labs to change how they do things.
33.
by trollbridge 18d ago
My website’s TOS says not to use it to train AI without permission, yet my website is in the training set of all the big models. So… my TOS doesn’t matter, but theirs does?
34.
by trollbridge 19d ago
It would also make testing far easier. I got bit with something that failed on real life iPhones with language set to es_US that worked fine on the simulator.
35.
by trollbridge 19d ago
Any competent graphic designer can do this. I think it’s a lack of effort because… they don’t need to spend the extra $100.
36.
by trollbridge 19d ago
This is genius - I’ve had good look using AI to identify fakes (or just reproductions) in other areas. It’s surprisingly good, and works fast enough to be usable in a fast paced live auction environment.
37.
by trollbridge 20d ago
Well, let's just say that Amex and Visa have been doing advanced fraud detection, using ML techniques, for years... and yes, some of that inside the z/TPF environment.
38.
by trollbridge 20d ago
I'm very un-motivated to do so since then I'd have more work to do, and I'd be making less money.
39.
by trollbridge 20d ago
No, but the 3 of us can now accomplish a lot more than we used to. We spend around $450 a month on hosted inference plus capital / utility costs of around $100 a month in local inference. And our 2 non-technical staff are now busy desi
40.
by trollbridge 20d ago
It’s a rebranding of ML which IBM has promoted for quite some time, used in a rather conventional fashion. AI does not always mean LLM.
41.
by trollbridge 21d ago
My biggest question is how I would open a claim. I can’t find a hotline number to call to open a claim (or to give to someone else who thinks they have a claim against me).
42.
by trollbridge 21d ago
I'm finding Corgi so far to be really difficult to understand. The sales reps (agents?) you work with don't seem to understand common insurance industry terms, like me asking "is that actual cash value or replacement cost?&qu
43.
by trollbridge 21d ago
Thank you - great response; so in other words, you can insure people that Corgi (and lots of other insurers) would eventually say "I'm sorry, but we just can't figure out how to insure you".
44.
by trollbridge 21d ago
1 trillion divided by 8.5 billion is $117. Global GDP is around $126T, so accumulated wealth of less than 1% of global GDP can barely move the needle.
45.
by trollbridge 21d ago
How is this different than Corgi?
46.
by trollbridge 21d ago
It’s not “de growth”. I want sustainable growth, which to me means: - no NDAs signed with elected officials who then stonewall the public and hide behind them - paying the same prices for water and electricity that I pay - paying the same
47.
by trollbridge 21d ago
Plenty of ice cold water in northern Quebec.
48.
by trollbridge 21d ago
Plenty of power in northern Quebec.
49.
by trollbridge 22d ago
The rare instance I see them is when needing to use the actual Facebook phone app, and then I see the ads targeted at me. Basically wall to wall ads for Cursor, Anthropic, and Littlebird.
50.
by trollbridge 22d ago
Based on the ads flooding me on LinkedIn, Facebook, and anywhere else I see ads, Cursor is really going after this market right now.
51.
by trollbridge 22d ago
In other words, plausible deniability. “We aren’t watermarking your text.” (Today, and we’ll start tomorrow.)
52.
by trollbridge 22d ago
Probably just means they haven't (yet) figured out how to do the watermarking carrying enough bits of data to encode a GUID.
53.
by trollbridge 22d ago
I've had my thinking in this area shift, too; I now believe China (and a few other places, like Vietnam) have "won" at manufacturing because their combination of government, cultural values, and the masses of their people fro
54.
by trollbridge 23d ago
Accelerando remains my favourite.
55.
by trollbridge 24d ago
Meanwhile, other AI firms are thirsty for data - Microsoft could have just slipped in a terms of service change to allow training (it’s sort of in there already), and presto, exabytes of data to train on, with full metadata so you even know
56.
by trollbridge 24d ago
Or AWS, or Microsoft, etc who have all the other stuff enterprises want.
57.
by trollbridge 24d ago
Abliteratuon sloghtly reduces the strength of the model - in my opinion it’s around the same jump as going from 5 bit to 4 bit.
58.
by trollbridge 24d ago
They’re very nice for some things. One challenge I run into is I run many agents at once, so local resource are a tiny fraction of the total inference we’re using. Every developer with his own Pro 20x, Max, etc. accounts is letting us hit l
59.
by trollbridge 24d ago
My M5 Pro gets around 12-15 (6 bit MTP), although I haven’t worked on optimising it at all yet. A nice thing about running locally is you can run an uncensored model and you don’t have to worry about TOS violations on your OpenAI account wh
60.
by trollbridge 25d ago
Building Android is heavy, and so is Chrome.

More ›