KD

@sama May we get limit reset plsss

artificial goblin intelligence achieved

@ayo32990345 Infinity Stones???? I normally like AI, in fact, I think it's very promising for animation, but this is just slop.

Now we know what will happen if superman fought conquest

@OppenOm @LuzAscendente @BlackBondPtv For coding xai not there yet but image to image and normal day usage is better grok is more daily informed also can disagre with u.Eventhis morning gpt try to delete and recode all projekt for simple request insted of saying it cant be done this way

@LuzAscendente @BlackBondPtv Tu rêves ?

🚨💣⚖️JOUR 2 DU PROCÈS MUSK vs OPENAI : ELON LÂCHE DES BOMBES ! Pendant 3,5 heures de témoignage, Elon Musk balance tout sur ce qu’il considère comme une trahison d’OpenAI : Il a injecté 38 millions de dollars dans ce qui devait rester une organisation non lucrative et open source. Ils lui avaient même promis des actions. Mais il refusait : « Une œuvre de charité ne devrait pas distribuer des actions, ça sent le pot-de-vin. » Quand Microsoft a injecté 10 milliards, Musk envoie un SMS à Sam Altman : « C’est quoi ce bordel ? D’où vient tout cet argent ? » Réponse de Sam : « Ouais, ça a l’air vraiment mauvais… Mais si tu veux des parts , on te les donne. » Résultat : Musk réalise qu’il a financé une « œuvre de charité » qui s’est transformée en machine à cash pour Sam Altman et ses associés, avec une introduction en bourse en ligne de mire. Et cerise sur le gâteau : les avocats d’OpenAI ont tenté de rabaisser Grok en disant qu’il était loin derrière ChatGPT. Musk a répliqué cash : « Plus maintenant. » Rendez-vous pour la suite demain.

@sama If someone is using agentic feature, pls ensure they can always access model from the same provider because after Claude switched to multiple providers, it became apparent that models working across architectures understood each other less.Token usage increased,quality decreased

we have updated our partnership with microsoft. microsoft will remain our primary cloud partner, but we are now able to make our products and services available across all clouds. will continue to provide them with models and products until 2032, and a revenue share through 2030.

@welovesweetboy_ He has bug brain show confermed when he try to solo trag.

oliver went through worse trauma but i bet he won’t have ptsd every 5 seconds like this performative guy mark did lol

@sama Idk man openai decline my card when its free but orher times has no problem with taking money...

We want you to have a lot of AI!

@e_fosa08 @SymplyTonbra I think armor able to replicate food creator from ship because it just created fking ps5

@SymplyTonbra Because I wonder how long she played games without food

Zoe is still a human so what was she surviving on for food 🤔 ?

@crownemmanuel @MatthewBerman Bro its normal to get difrent responses from llms its how it works random seed + temperature settings is most common reason for this

@MatthewBerman I think this video might be staged. He might have embedded some system prompt or instructed the model earlier to behave this way. When I ask the same question on chat voice mode, I get a different response.

How does he find these

@MatthewBerman GPT just doesnt like u bro.

@Truthoverclout @ThomasConnorsJr @DailyMail @Synthid Bros prof cant tell how many r in strawberry

@ThomasConnorsJr @DailyMail The 2nd image has been edited by Google AI (Nano Banana) and it has a SynthID, the digital watermark for AI content that Google embeds in all its AI Images. You can test it by uploading this image to Google Gemini and typing @SynthID. support.google.com/gemini?p=synth…

You are disgusting @DailyMail (Daily Mail busted doctoring photo). Even after his death, the media still does this shit to MJ

@albaybencerich @ataaberrk Andaval ss'den tabletinin dockunu kırpamıyor ama ülke kuran adama laf ediyor xD

@ethaleu @iflzi @BikiniBodhi @ohnePixel Its tos breach

This is the last I'll say on this because holy fuck @ohnePixel is actually just clipfarming and it's annoying. Here's the proof the discord account related to all the cheats are his, these are still available if you go to the website. Which is also the same discord tag Mooda used to promote in his old videos, still up here: Both #Rawrri5333. youtube.com/watch?v=SQFxX6…

@Jaereku @ElijahSchacher @ramennnnoodle Nah anime is fine they gonna ban games prob they say he seen this kind of behevior from pubg cs ....

@ElijahSchacher @ramennnnoodle Literally my country dawg. It’s so fucking over for Turkish Re:Zero fans

worst Natsuki Subaru mention of all time

@XxJakez123xX No one at tukiye even pointed animes yet they saying he did it because infulance of bad games. But child has known issues it has nothing to do with rezero + who the fuck lets his piscologicly unstable child to police level gun training he was 13

Re: zero being blamed for a school shooting?

@DHL1026746 @BikiniBodhi "No proof" :D

@BikiniBodhi So I guess we are going to keep running other peoples names in to the ground with no proof just to make a bag, biggest loser in this community and everyone agrees with it.

”falsely banned”

🏆 WALLHACK x Wooting Collab Giveaway 1x SP-005 Ghost in the Shell 1x Wooting 80HE Ghost Steps to enter: - Follow @WALLHACKHQ + @WootingKB - Like and RT the post - Tag one friend 🌎Global Giveaway 🗓️Ends April 20th ※ You can also participate via Instagram!

@AlicanKiraz0 Tebrik ediyorum elinize sağlık. Gelecekteki modellerde sansürüde kaldırırsanız harika olur modeller çok fazla şunu yapamam buna iznim yok demeye başladı.

Selamlar Dostlar, açık kaynak olarak paylaştığım yeni Türkçe muhakeme modelim Kızagan-E4B yayında! 🎉🙏🏻 Kızagan-E4B'yi, Google DeepMind'ın iki hafta önce yayınladığı Gemma 4 E4B-it base modeli üzerine, taxonomy ve domain dağılımını oluşturduğum 540 bin satırlık (~2 milyar token) çok eksenli Türkçe SFT veri kümesi ile fine-tune ederek oluşturdum. Veri kümesini; 8 alan × 3 zorluk × 5 üslup × 9 persona × 8 çıktı formatı × 7 beceri ailesi (25+ alt beceri) eksenli bir taksonomi üzerine inşa ettim — talimat takibi, yeniden yazım/özetleme, yapılandırılmış çıkarım, kaynağa dayalı soru-cevap, günlük ve matematiksel muhakeme, Türkçe dil hassasiyeti (yazım, noktalama, morfoloji, kayıt) ve güvenlik/belirsizlik başlıkları ağırlıklı yer tuttu. Base modele (google/gemma-4-E4B-it) karşı benchmark sonuçları: 📊 Türkçe MMLU (malhajar/mmlu_tr-v0.2): %30.42 → %33.18 (+2.76 / +372 doğru) 🧮 Turkish GSM8K (strict): 14.06 → 35.94 (+21.88 puan — yaklaşık 2.5× sıçrama) Modeli 5 farklı biçimde yayınladım: 🔹 Base (BF16) — Transformers / vLLM 🔹 Q8_0 GGUF — llama.cpp, Ollama, LM Studio 🔹 MLX 8-bit / 4-bit / fp16 — Apple Silicon için Türkçe açık kaynak NLP ekosistemine küçük de olsa bir katkı olmasını; özellikle veri güvenliği nedeniyle çevrimdışı çalışması gereken, Türkçe muhakeme ve yapılandırılmış çıktı gerektiren yerel kullanım senaryolarında topluluğa umarım faydalı olur. 🙏

@techolay Tarihin en kötü performansı ile çalıştırdı sadece llm için kullanılabiliriyor tek numarası apple onaylı ve gelişmeye önü açık olması daha gelişip adapte olursa belki ilerde PP egpu TG apple gpu da fln yapılabilir olur

Apple Onaylı TinyGPU, Mac’te RTX 5090’ı Çalıştırdı techolay.net/rtx-5090-macte…

@AlainMargerie77 @wono_strategy If poli cancel something money is not going back insted every party takes equal so lowest percentage get highest money its a scam

@wono_strategy @wono_strategy Why?

Not gonna end well

@SenValeTV May i have one pack pls

Ubisoft has kindly sent me codes for the #R6Redacted event packs! I will be giving away on stream randomly throughout the day on my channel! Appreciate you Ubisoft ♥️