Anecdotally, I suddenly had almost my entire friend group on matrix after that as well. I had campaigned for months but AV was the thing that got them to finally sign up.
Doesn't the way OAI's proof is formulated essentially rule that out, in that as long as an alternative solution concerns option C, it'd have to live in this same solution space they carved out?
I have seen 30k line cpp files even a decade ago (World of Warcraft server emulator, gameplay logic of a boss enemy), and was told it is fairly normal in large software (even 100K not being unheard of), so I'm not sure if it's that much of an LLM thing.
Is that why session compaction stopped working for me in VS Code at the end of this week, or is that just the integrated extension itself having a normal one?
Though it's the same extension that can't keep its session timestamps straight, randomly hides sessions I was just in (then suddenly remembers them after going in and out of a session), and completely shits itself visually when using OpenAI's models, so maybe it really is just the latter.
Was there any TLS interception in place? Mind you, that can be detected and ignored...
Cause the (obviously AI generated) page is fairly ambiguous in this regard. In one section it claims certain pieces of info were sent outright. In others, it refers to them as "Privacy Threat Vector" items. Were they possibly sent or were they actually sent? Why is this left unclear? Why is transmission alone counted as evidence of later misuse? What data is technically necessary to send as part of a protocol?
I'm really quite tired of the run of the mill "privacy minded" folks thinking they're the hot shit because they can launch WireShark, and gawk at packets flying about. Like no, various corporate SNIs appearing in a chatty network log is not evidence for illegal or unethical corporate espionage/surveillance, especially not a clear one. Nor is the network log being chatty any evidence one way or another. Do you really think that surveillence is a more likely explanation for them than just regular enterprise sprawl?
If you're bringing receipts, bring them whole, disclaimers and limitations included. Any analysis that stops before decrypting the traffic is deeply unserious, and only serves to discredit actual research findings & real violations of privacy.
Have you considered making the model leverage computer use via Work/Codex?
I'd imagine making the agents use the same tools that people utilize would greatly dial back how AI-generated these things look, and it would provide much better control too.
With Astra being genuinely much better at this, it should be more viable than ever.
If you break out more sophisticated tools - give Claude Code some skills, let it work in SVG or PDF, create bitmap assets and composite them with real fonts and vector assets, I'm sure you can achieve a lot.
But the point of this blog post was to demonstrate what a PTA mum can achieve even with just ChatGPT, if they just prompt a little more creatively.
The proviso though is that you don't get many tries on free tier. Or if you pay for more tokens, you can fall into a rabbit hole of iterating where it's no longer a quick task.
Mathematically speaking, 18 / 3.6 isn’t “reducing” by 3.6X, it’s “dividing” by 3.6X. Reducing would be 18 - (18 * 3.6), which is obviously wrong. By your formula, “reducing by 50%” would be 18 / 0.5, also obviously wrong.
Yes, people do say things like “reduce by 3.6X” and are understood to mean what you said, but they also say “literally” when they mean “figuratively”. It doesn’t bother me but I can understand why math oriented people would be annoyed, and I personally would never say “reduced by 3.6X”, but instead “reduced by 72.2%”.
I was sincere, because as you point out, the phrasing is not actually ambiguous here. There is only one way to interpret this that is coherent and sensible. The usual % shenanigans weren't even on the table for me.
Not that I'd have ever seen "reduced by 0.5x" or any other value below 1x, probably for this very reason. What I do see is "reduced to 0.5x", in which case you're supposed to swap the division for multiplication.
Percentages on the other hand are a whole another can of worms, even if these forms are principally interchangeable, and I find them a lot more confusing a lot more often.
Not that this would explain the whole mean/geomean thing.
How often should we allow human provided targets to go unverified by a human before killing every innocent student inside the target?
That is, the question of whether targets should be verified is orthogonal to the question of whether humans or AIs supply targeting data with fewer errors.
Not often, and when it does happen, those that commit criminal negligence should be served justice.
It shouldn’t be hard to gather intelligence that shows “kids go to school here.”
Then again, bad American intelligence dragged them into a very expensive war in Iraq on false grounds and we’re still trying to get justice served 23 years later.
That we do. So it'd be pretty cool if that was done responsibly too, and being able to answer my question very much matters for that.
If these things do outperform servicemen, or if there's not enough data to assess that, then selectively reporting this in headlines is not exactly helping anyone, quite the contrary.
Especially knowing that there's a significant anti-AI sentiment among people as-is, I really wouldn't put it behind news outlets to couple that with some conveniently missing context and take advantage of people for some cheap clicks. Kinda been the theme for a while now if you noticed.
Whether that then results in society making responsible decisions, and holding the correct people accountable...
I dread the day I'll make claims like this, even figuratively.
Even at a most surface level reading, this would suggest something like their MAU (monthly active users) at least halving soon/already.
I'd be very surprised if it was trending down even, let alone at this! Forget hyperbole, this is an outright überbole.
By all means though, got any data?
reply