jjani

13 ms·

181.
by jjani 1y ago
Is it really this easy now to get your article high on HN with 100 comments? The findings are completely meaningless. "Agenticness" depends so much on the specific tooling (harness) and system prompts. It mentions Copilot - did it
182.
by jjani 1y ago
Neither will be. Both OpenRouter and Azure (through requiring and enterprise agreement, only available to large orgs with 500+ devices) require it for o3 to this very day, and already do so for GPT-5, the main model under discussion in this
183.
by jjani 1y ago
It does for the model this thread is about: openai/gpt-5.
184.
by jjani 1y ago
It's only a video face scan and your legal ID to SamA, what could possibly go wrong
185.
by jjani 1y ago
They first introduced it 4 months ago. Back then I saw several people saying "soon it will be all of the providers". We're 4 months later, a century in LLM land, and it's the opposite. Not a single other model provider a
186.
by jjani 1y ago
Yup! Oh plus a video face scan, I forgot to mention.
187.
by jjani 1y ago
Yup! Nice play to get a picture of every API user's legal ID - deprecating all models that aren't locked behind submitting one. And yep, GPT-5 does require this.
188.
by jjani 1y ago
On average, they sure put out better customer service than the likes of Google and Meta.
189.
by jjani 1y ago
> Web app which takes 10+ sec to fully load? That's ok, focus on the new features! This web app wouldn't survive unless it already had backing of the Microsoft level or did offer immense unique value to its users.
190.
by jjani 1y ago
And this is exactly why they need to be broken up. If that offering was their core service, you can bet it would be the priority of the middle managers.
191.
by jjani 1y ago
This is what I was referring to, and we already have seen this happen in the wild with PayPal at one point (possibly still) blocking passkeys from e.g. Firefox. For now the argument against this seems to be that "Apple zeroes this out
192.
by jjani 1y ago
> So I guess the question is, would people be willing to pay for good support? They're already doing so, just not receiving it. Microsoft made a yearly profit of 101 billion dollars , or better phrased, 101,000 million dollars . T
193.
by jjani 1y ago
Even with backups, the attestation issue makes them awful.
194.
by jjani 1y ago
It can be just as true for LLM output. One can use LLMs simply in a way to gather information, replacing Google by a much better one, and create lesson plans yourself based on the information provided. Nothing stands in the way of you doing
195.
by jjani 1y ago
> There are plenty of places to place the blame, but this is a function of any LLM, and a funcamental way LLMs work, not just a problem created and profited from by Google- for example the open-weight models, where no-one is actually pro
196.
by jjani 1y ago
Seemingly infinite, I don't think they've invited anyone from the list so far.
197.
by jjani 1y ago
Nowadays the biggest EU parties are ones that are effectively against the EU, so they're using the most effective way to kill it: make their citizens hate it by passing such laws. If you create a diagram of "pro-EU" and &quot
198.
by jjani 1y ago
Haha, got a laugh out of me assuming you were intentionally demonstrating by example :)
199.
by jjani 1y ago
Can confirm it's "Not just the GPT's - it's all of the frontier models." who are addicted to that one. IME the only reliable way around it when using an LLM to create blog-like content is to have actual hard lists o
200.
by jjani 1y ago
I've done years of private 1:1 teaching and some class teaching though not class lecturing, which is presumably the material you're talking about. > As a former teacher, I know you need to have a good grasp of the material you
201.
by jjani 1y ago
"drug x cures y in humans - without causing z where bad_y - bad_z < ε".
202.
by jjani 1y ago
Typical how the likes of the FT keep pushing these articles en masse about such a small event, when the opposite which is magnitudes more widespread and keeps destabilizing societies more and more, is not. It's all PR, entrenched inter
203.
by jjani 1y ago
All those teachers should indeed be banned from using AI. But that's not because LLMs are incapable of the things they're using them for, in a way that would be an improvement over how those same teachers were doing those tasks pr
204.
by jjani 1y ago
On the other hand, they have always exposed their raw chain of thought, so you know exactly what you're paying for, unlike OpenAI who hides it. Similarly they allow an actual thinking budget rather than vague "low, medium, high&qu
205.
by jjani 1y ago
This concept is very far from all rainbows and sunshine [1], exactly what you'd expect to happen leaving it all up to the market indeed. [1] https://www.bbc.com/news/articles/ce9jv1xyxm0o
206.
by jjani 1y ago
https://www.munhwa.com/article/11487922 But really, being deeply ingrained into society and having personally worked at such places here in Korea is worth more than flawed statistics, which really all statistics are
207.
by jjani 1y ago
The situation in Korea is completely different. There are tens of thousands of jobs available, including white collar, but people don't want to work for those companies - everyone goes for the exact same handful of jobs, the very top.
208.
by jjani 1y ago
Most likely it's both real and LLM generated. As in they gave their case, asked the LLM to write this article, did some very minimal editing and copy-pasted.
209.
by jjani 1y ago
To the latter: Absolutely all of it, though I put my money on Claude, it has more of its prominent patterns.
210.
by jjani 1y ago
There's just so many tells in this one though and they aren'tn new ones. Like a dozen+, besides just the entire writing style being one, permeating through every word. I'm also pretty shocked how HNers don't seem to noti

More ›