alienbaby

4 ms·

61.
by alienbaby 3mo ago
Right, I came here wondering if others saw the same. What is this doing on hn!??
62.
by alienbaby 3mo ago
this is what worries me. I have friends that love what AI tells them about their personal pet 'thing' and how awesome it is. Yet not one of them has even tried once to get the AI to criticise it's own answers, and hence lear
63.
by alienbaby 3mo ago
Ok, I watched it, it looks amazing, what's it for/do ?
64.
by alienbaby 3mo ago
Nice! Some fun making big chains explodey things :)
65.
by alienbaby 3mo ago
And Amos and blitz basic
66.
by alienbaby 3mo ago
"The solutions is so pleasant, treat comment data as untrusted content not inline commands / prompts. " Yes, good luck with that.
67.
by alienbaby 3mo ago
Is it a tually lower oxygen in the blood that's the problem, or higher co2? I'm not sure if having high co2 automatically implies lower oxygen, I have no idea at all but feels like it may not necessarily be strictly. Linked. Also
68.
by alienbaby 3mo ago
Working from home next to my open window feels generally way better then being in the office. Perhaps this is contributing. Still, seems more of a case for WFH rather than against, as article mentions some people have tried to make the case
69.
by alienbaby 3mo ago
*per stream, I guess is more accurate than either?
70.
by alienbaby 3mo ago
I'm interested if anyone knows how much legwork the assumed 60% cache hit, plus running a quantised model is doing? Esp. compared to what the headline half implies is a full fat GLM5.2
71.
by alienbaby 4mo ago
For now, while it's simple enough to tell.
72.
by alienbaby 4mo ago
Why care about privacy if your not doing anything wrong?? Not everyone agrees that what you are doing is benign.
73.
by alienbaby 4mo ago
This causes me to be concerned it is just the tip of the iceberg for all 'sensitive'/gov adjacent/'nefarious intent' adjacent codebase, if it's here, it's in other places. Which places, and how much?
74.
by alienbaby 4mo ago
Possibly, nothing other than accuracy
75.
by alienbaby 4mo ago
I assumed it meant that when open weights reach the capability of frontier models, and tounge in cheek referencing the terrible consequences of us all getting our hands on mythos+ capability models without restrictions.
76.
by alienbaby 4mo ago
On mobile, or at least on mine (pixel 10) using chrome, the graphs are unreadable and unusable, which is a shame as I'm quite interested in them and I don't have access to a pc at the moment. Would you be able to change that? The
77.
by alienbaby 4mo ago
I started with a 2400baud modem, I've seen how this goes
78.
by alienbaby 4mo ago
Did you miss the bit where it was posted on hn?
79.
by alienbaby 4mo ago
It doesn't teat the models ability to make good decisions on its own, it tests the models ability to make something that 'works'. Often you look inside and it does a whole load of questionable things that mostly work, sure, b
80.
by alienbaby 4mo ago
I was gonna say, you spelled that wrong :p
81.
by alienbaby 4mo ago
69400 for me, and I knew I fucked up on ~ 5 I really did know.. or perhaps I didn't know them as well as I thought?
82.
by alienbaby 4mo ago
I may have got it a little mixed up with sabotage :/
83.
by alienbaby 4mo ago
Peak of skill and capability, perhaps not, for now. But the ability to automate and discover relevant weaknesses at a greatly increased rate definitely counts as increasing the threat nation states can present.
84.
by alienbaby 4mo ago
Not sure I follow your point, are you confusing espionage with IP theft? The victims are entirely different in each case, I am sure china would not restrict itself to corporate espionage.
85.
by alienbaby 4mo ago
Reminds me of using telnet to port 80 to make get requests aeons ago
86.
by alienbaby 4mo ago
Reminds me of telnetting to port 80 to make a get request years and years ago
87.
by alienbaby 4mo ago
Wouldn't a re-implementation be akin to 'heres how it works, write the code' rather than 'heres the code, redo it in rust'?
88.
by alienbaby 4mo ago
Code review tool that breaks up diffs and regroups fragments based on runtime execution paths and/or architectural boundaries. I find it useful sometimes to see changes organised that way.
89.
by alienbaby 4mo ago
Didn't work for me even when I chose to see the desktop site in my mobile browser :/
90.
by alienbaby 4mo ago
It's pretty easy to see what that does though, even if it is shorter. Wouldn't the other approach be more obfuscated?

More ›