mathstodon.xyz is one of the many independent Mastodon servers you can use to participate in the fediverse.
A Mastodon instance for maths people. We have LaTeX rendering in the web interface!

Server stats:

2.8K
active users

#O4

3 posts3 participants0 posts today

#Unicode:

#ChatGPT hinterlässt #unsichtbare #Zeichen im #Text

In den neueren #Sprachmodellen von #OpenAI hinterlässt die KI offenbar absichtlich unsichtbare Zeichen.

Dem Bericht nach hinterlassen die Modelle #GPT- #o3 und #o4-mini unsichtbare #Unicode-Zeichen in dem generierten Text. #OpenAI kündigte diese Maßnahme bisher nicht an.

golem.de/news/wasserzeichen-ch

Golem.de · Wasserzeichen: ChatGPT hinterlässt unsichtbare Zeichen im Text - Golem.deBy Andreas Donath

The AI has lost its marbles...
...signoff message.
A 4 hour session to try to write a complex script in #bash.

What began as a simple exercise, became a "why are you so dumb" trial.

TLDR; #Chatgpt #o4 don't do BAsH gud.

However the new "Research" mode of ChatGpt is impressive, it took 20 minutes realtime.

Vibey (Worker) comparison
between #o4 #Chatgpt and #Claude Sonnet 3.7

So recently I got a new CC and had difficulty getting it in #Antrophic. Because I have grown reliant on the PRO model in my daily. I paid the #AI tax to #OpenAI.

Here is my experience.

1. I'll restate this because it needs restating. The free models are dumber. The only meaningful assessment can come from the pay-for model.

2. AI moves at breakneck speed a month in AI is worth at least 6 elsewhere. Would you believe there are still 6-finger jokes floating around, even though current pro Gens done that for a year+.

3. The new ChatGpt model definitely seems smarter.
It seems to unnecessarily burn compute though, offering multiple solutions to issues.
I liked how it quickly adapted it's persona to my work style.

4. I like the new "vibe coding" refactoring, where it will go line by line through the code changing it. Very SciFi.

5. The new Pro sub for OpenAi comes with Gen subs (value+) so you can create images (Anthropic doesn't have that).
Also #Sora sub so you can make 10s videos, if you have seen Sora videos, they are mind-blowing.

7. It has another model called "Monday" which just works like an asshole prompt. Another proof that most users still have a lot of ground to cover in #promptengineering

Overall, I think PRO ChatGpt is slightly better than Claude, though I have gotten used to Claude.

Braucht jemand noch etwas Lesestoff zum Samstagsfrühstück? 📰 🍵🥐☕

„Detektoren von Einsteins Wellen in den Startlöchern“, ein Online-Artikel vom „Sterne und Weltraum“-Chefredakteur über den Start der zweiten Hälfte des 4. Beobachtungslaufs und die Erwartungen an die Beobachtungskampagne.

➡️ spektrum.de/news/gravitationsw

Spektrum.de · Detektoren von Einsteins Wellen in den StartlöchernBy Andreas Müller

🚨 Rätselhafter Lückenfüller – Bereits kurz nach dem Beginn des 4. Beobachtungslaufs haben die @LIGO@astrodon.social-Virgo-KAGRA-Kollaborationen ein überraschendes Gravitationswellen-Signal beobachtet. 😲

Es stammt von der Verschmelzung eines Neutronensterns mit einem unbekannten kompakten Objekt, wohl einem ungewöhnlich leichten Schwarzen Loch, das in der sogenannten „untere Massenlücke“ zwischen den schwersten Neutronensternen und den leichtesten Schwarzen Löchern liegt.

ℹ️ aei.mpg.de/1138268/mysterious-

🚨 Mysterious object in the gap – Shortly after the start of the fourth observing run, the @LIGO@astrodon.social-Virgo-KAGRA collaborations detected a remarkable gravitational-wave signal. 😲

It's from the merger of a neutron star with an unknown compact object, most likely an unusually light-weight black hole falling into the “lower mass gap” between the heaviest neutron stars and the lightest black holes.

ℹ️ aei.mpg.de/1138125/mysterious-