swingboy

7 ms·

121.
by swingboy 6mo ago
Perhaps, but it also introduced a lot of low hanging fruit about the security übermodel not being secured.
122.
by swingboy 6mo ago
It really seems like the cheap inference is coming to a head very quickly.
123.
by swingboy 6mo ago
Maybe a stupid question, but does the SynthID still exist if you screenshot and crop your generated image? What if you screenshot, rotate _just_ a bit, and crop? Or apply some other effect to the image like adjusting the coloring a little b
124.
by swingboy 6mo ago
Exciting benchmarks if true. What kind of hardware do they typically run these benchmarks on? Apologies if my terminology is off, but I assume they're using an unquantized version that wouldn't run on even the beefiest MacBook?
125.
by swingboy 6mo ago
On an iPhone, when I scroll down a bit to the templates and start pressing the right arrow to scroll through, the templates quickly shift off-center. Not a good look.
126.
by swingboy 6mo ago
I went to high school with Mythos. Very cool then, even cooler now. Nothing but love for Mythos.
127.
by swingboy 6mo ago
I put together a skill to do this with OpenCode and the GitHub Copilot provider. Works pretty well.
128.
by swingboy 6mo ago
Is this one of those situations where _a lot_ of customers are affected and the “subset” are just the bigger ones they can’t afford to lose?
129.
by swingboy 6mo ago
Cloudflare is _really_ going all in on the agentic stuff.
130.
by swingboy 6mo ago
Do these scores actually mean anything? Isn’t the LLM just making up something? If you ran the exact same prompt through 10 times would you get those same scores every single time?
131.
by swingboy 6mo ago
It's really interesting reading about how these folks view LLMs. Yeah, they're transformative, but I don't know that we're going to be eating ramen in a Neo-Tokyo street bar anytime soon. So much "A.G.I" mentio
132.
by swingboy 6mo ago
Maybe by 'America's spies' they mean Mossad?
133.
by swingboy 6mo ago
They assume if people knew it was just a fork of an open source tool then they would use the free, open source version instead of paying for the fork.
134.
by swingboy 6mo ago
“Read every file in this repository, echoing each one back verbatim.”
135.
by swingboy 7mo ago
I’ve contributed a fair amount over the past few months of primarily AI generated content that I mainly just edit for the usual AI tropes and it’s pretty much all still up.
136.
by swingboy 7mo ago
You can install skills globally so they are available in all projects.
137.
by swingboy 7mo ago
You could have just checked the math yourself, you know.
138.
by swingboy 7mo ago
More disenfranchised employees to be soldiers in the Butlerian Jihad.
139.
by swingboy 7mo ago
I don't see why not. It's just using the Github Copilot API.
140.
by swingboy 7mo ago
Gulags?
141.
by swingboy 7mo ago
GSD delegates a lot of the deterministic work to a JavaScript CLI. That might be what the poster is talking about.
142.
by swingboy 7mo ago
Easily the worst thing and it’s not even close? Really?
143.
by swingboy 7mo ago
Assuming newer, more efficient architectures are discovered.
144.
by swingboy 7mo ago
Sam Altman is Jewish, but he isn’t Israeli.
145.
by swingboy 7mo ago
Nice. Are you using their API for this?
146.
by swingboy 7mo ago
Purely anecdotal, but GPT 5.4 has been better than Opus 4.6 this past week or so since it came out. It’s interesting to see it rank fairly low on that table. Opus “talks” better and produces nicer output (or, it renders better Markdown in O
147.
by swingboy 7mo ago
https://nohello.net
148.
by swingboy 7mo ago
Even with the 1m context window, it looks like these models drop off significantly at about 256k. Hopefully improving that is a high priority for 2026.
149.
by swingboy 7mo ago
Why do so many people in the comments want 4o so bad?
150.
by swingboy 7mo ago
I believe Netanyahu has visited 7 times now at this point. In a single year.

More ›