HackerTrans
TopNewTrendsCommentsPastAskShowJobs

scrollop

420 karmajoined 6 lat temu

Submissions

Deep research benchmark shows how poor LLMs are at writing accurate reports

youtube.com
3 points·by scrollop·9 miesięcy temu·1 comments

comments

scrollop
·5 dni temu·discuss
I thought that he invented the computer chip?
scrollop
·7 dni temu·discuss
Interesting.

Perhaps using non-ozone negative ionizers would also help.
scrollop
·10 dni temu·discuss
"I think the models are being optimized for wealth extraction from users and companies, instead of solving problems."

YES! They introduced the new tokenizer to increase token generation by upto 33%.

On top of this, Anthropic are generating almost twice as much revenue per paid user than openai - whilst their subscriptions have lower usage limits than openai's:

https://youtu.be/gK-7TKC7kvY?si=kx0qPE1rw-UCI-Jn&t=650
scrollop
·13 dni temu·discuss
We order ultrasounds all the time for shoulders (for like soft tissue issues; for trauma, you'd start with an xray). For other joints, such as the knee, MRIs are a better choice (unless htere has been substantial trauma, in which case xray initially or further), though more expensive, unless you're excluding a Baker's cyst, in which case an ultrasound is fine.

Since MRIs are more expensive, private doctor's might order them instead of an ultrasounds.

(I'm a doctor)
scrollop
·17 dni temu·discuss
Perhaps they used gpt 5.5 mini to draft emails. Create a coffee schedule.
scrollop
·17 dni temu·discuss
Seems it has started already:

https://www.reddit.com/r/ClaudeAI/comments/1ucu6og/any_solut...
scrollop
·18 dni temu·discuss
Interesting beside the time when Anthropic were banned for not allowing the MoW to use claude to kill people, then OpenAI come in and swoop up the contract.
scrollop
·19 dni temu·discuss
Pranayama is translated more as "control of the breath"- there are some techniques that are much faster than normal respiration eg Bhastrika and Kapalbhati.
scrollop
·24 dni temu·discuss
Would be interesting to see where gpt 5.5 pro extended is.
scrollop
·24 dni temu·discuss
And these same people likely fund "reports" and "news" with misinformation to make it confusing for the average person.
scrollop
·w zeszłym miesiącu·discuss
https://artificialanalysis.ai/evaluations/omniscience
scrollop
·w zeszłym miesiącu·discuss
I'd recommend carefully looking at a few benchmarks (even though generally relying on benchmarks is problematic)

https://artificialanalysis.ai/evaluations/omniscience

Esp check the Hallucination rate for Deepseek - it's not good.
scrollop
·w zeszłym miesiącu·discuss
Imagine if cutting edge AI companies could decide to using their world best AI to

1) Develop software for linux

2) Provide decent support
scrollop
·w zeszłym miesiącu·discuss
Out of interest - do you trust google reading all your emails? What do you think about privacy?
scrollop
·w zeszłym miesiącu·discuss
I find it odd how so many tech involved people here use gmail - are privacy concerns not a concern for them?

I moved to mailbox.org years ago. Pay a few pounds a year for private email with webtools and drive and don't have google snooping my emails and sending me targeted ads.
scrollop
·w zeszłym miesiącu·discuss
Agreed - 5% fees is quite high considering the volumes of tokens involved.
scrollop
·w zeszłym miesiącu·discuss
Though you pay 5% fees? Not worth it for me with the volume of tokens used.
scrollop
·2 miesiące temu·discuss
Also increased, or we're now aware of, higher rates in long distance runners.
scrollop
·2 miesiące temu·discuss
You should try hosting it yourself in docker. Absurdly easy to do if you get an llm to do it and it works very, very well.

Hope they don't alter self hosting it.
scrollop
·2 miesiące temu·discuss
Since it appears that LLMs can't achieve AGI and lose hallucinations, I presume a new company will appear with a new architecture that can - what happens to the current behemoths and their stock prices? Will they jump architectures?

Splendidly interesting times.