Какая нейросеть лидирует в генерации текста сейчас? ChatGPT vs Gemini vs Claude
С каждой неделей наш мир пополняется новой нейросетью. Вспомните, каким был февраль, который принес нам новые модели от Claude, OpenAI и Gemini. Март перехватил эстафету, хотя эта неделя выдалась без особых новшеств. Небольшой клочок спокойствия за последний один-два месяца. В своей предыдущей статье я сравнил последние новинки от компаний, названных в абзаце выше. Сравнение происходило не во всех сферах, а затрагивало только программирование . В размышлениях я подумал: а почему не столкнуть модели в еще одном варианте задач? Речь идет о генерации текста , наверное, одно из наиболее популярных направлений при работе с нейросетями. Сегодня будет сравнение ChatGPT 5.4 , Gemini 3.1 Pro и, конечно же, Claude Opus 4.6 в задачах, которые связаны с текстом . Принимайте стратегически удобное положение, ну а я начинаю свое повествование.
https://habr.com/ru/companies/bothub/articles/1010212/
#Gemini #ChatGPT #Claude #openai #anthropic #Google #AI #ии #нейросеть
"Diktatorenartige Huldigung": Trump unterwirft die KI-Riesen
Семантический поиск vs полнотекстовый: сравниваем три embedding-модели на 10 000 категорий Ozon
Сравнил полнотекстовый поиск PostgreSQL ( tsvector / tsquery + GIN-индекс) с семантическим поиском через pgvector ( cosine distance ) на датасете из 10 019 товарных категорий Ozon. Три embedding-модели: GigaChat EmbeddingsGigaR (Сбер, 2560-мерные векторы, API) Qwen3-Embedding-0.6B (Alibaba, 1024-мерные, локальный инференс через HF Text Embeddings Inference на GPU) text-embedding-3-small (OpenAI, 1536-мерные, API) Прогнал 18 запросов в пяти категориях: синонимы и сленг, intent-запросы в свободной форме, подарочная тематика, cross-lingual (EN-запросы к RU-данным), абстрактные формулировки. Замерил латентность и top-5 с cosine similarity score. Разбор каждого запроса, таблицы и код — под катом.
https://habr.com/ru/articles/1010200/
#искусственный_интеллект #gigachat #openai #qwen #embeddings
RE: https://masto.ai/@teacherbuknoy/116219733025778304
Thinking about this is so insane. It's like saying "you can pay us to chew your food for you and we'll puke out a sludge that you can swallow for easier digestion."
Really wild stuff coming out of the University of Colorado from this OpenAI deal. Take this quote for instance:
«Before AI tools became ubiquitous, students and junior workers typically turned what they learned into artifacts—they would write a software function, develop a mathematical proof, draft an essay or sketch out a design... Now that AI can easily create artifacts, such outputs can no longer be considered the endpoint of mental work.»
This is how disconnected these people are from what academics, or anyone creative, actually do. I know a chatbot likely regurgitated this line, but someone chose to post it.
If that wasn't enough, OpenAI's president gave millions of dollars to Trump almost simultaneously with this deal going though at CU. It's absurdly easy to follow the money.
#Colorado #academia #CUBoulder #CU #Boulder #AI #OpenAI #USpol #research #noAI
El CEO de OpenAI describió un problema estructural del capitalismo frente a los inversores que financian ese mismo modelo. Sin propuesta de solución, sin mecanismo de transición laboral. Solo el diagnóstico. Que alguien con ese nivel de influencia lo diga en voz alta ya es la noticia.
https://www.glitchmental.com/2026/03/sam-altman-ia-trabajo-capital-2026.html
Widzi pan, panie Altman...
Gdybyście ChataGPT utrzymali jako open source a OpenAI byłaby naprawdę otwarta, to pan może nie byłby aż tak bogaty, ale wielu chętnie dołączałoby do rozwoju, jak i wspierało.
A tak to pleciesz pan bzdurę za bzdurą, wiedząc że czas OpenAI jest policzony i że albo etat w Microslopie albo Google, ewentualnie emerytura i zabawa w inwestowanie w startupy.
Nie jest mi pana szkoda.
https://spidersweb.pl/2026/03/sam-altman-ai-licznik-prad-woda.html
Interesting case of chatGPT bullshitting me in a new way:
Spotted an unusual bird on the balcony which I didn't know the name of so I snapped a photo.
Out of curiosity I sent the photo to ChatGPT and asked if it could identify the bird.
It said the name of some bird and remarked on how it is a common sight in Osaka.
Very strange, I am in Canberra.
That's when I realize that I have my VPN set to Japan.
So, clearly, chat-jippitty was using my IP location to try and narrow down its guesses.
(the guess was wrong, it was absolutely not a Japanese bird, still haven't figured it out the actual answer yet)
“We see a future where intelligence is a utility, like electricity or water, and people buy it from us on a meter.”—Sam Altman
Like we didn't see this coming. Like we didn't expect for this to happen like clockwork. Greedy white twinks and their capitalistic gluttony for money that isn't even real.
#AMD, #Broadcom, #Meta, #Microsoft, #NVIDIA and #OpenAI jointly announced today the formation of the Optical Compute Interconnect (OCI) Multi-Source Agreement (MSA) optical #scaleup consortium.
This consortium is focused on a multi-vendor, open ecosystem for optical scale-up interconnects. #AI, of course, being the big focus of the optical interconnects in 2026.
https://www.phoronix.com/news/Optical-Compute-Interconnect
Andromeda
Sora soll in ChatGPT integriert werden
OpenAIs Sora soll künftig in ChatGPT verfügbar sein. Das könnte dem Videogenerator Aufwind verpassen – aber auch Kosten erzeugen.
Lou LüederC'est sympa cette nouvelle feature de Firefox pour envoyer TOUT TON HISTORIQUE DE NAVIGATION à Google, Alibaba et OpenAI 🤡💩
https://www.omgubuntu.co.uk/2026/03/firefox-smart-window-hands-on
Oracle-Geschäft profitiert von KI-Boom
Oracle investiert massiv in KI-Rechenzentren. Trotz anfänglicher Zweifel an der Börse, ob sich die hohen Investitionen lohnen, brummt das Geschäft.
#Börse #ChatGPT #CloudComputing #KünstlicheIntelligenz #OpenAI #Oracle #Wirtschaft #news
Itamar Golan (@ItakGol)
트윗은 OpenAI가 'OpenClaw'를 인수했고 FaceBook이 'MoltBook'을 인수했다는 내용을 전하며, 회사명을 신중히 지으라는 농담으로 마무리합니다. 인수 소식은 AI 업계의 인력·기술 통합이나 제품 전략 변화 가능성을 시사합니다.
Wenn ich mir das mit dem ganzen KI Gedöns so recht überlege - es hätte mich nicht besser treffen können.
Ich bin 50 - Ich habe ausreichend Berufserfahrung um die Dinger steuern bzw. kontrollieren zu können und bin noch neugierig genug um den Umgang damit zu erlernen.
AMI Labs
🔸 Milliardaire, la jeune pousse d’IA de Yann LeCun s’installe à Montréal
https://www.lapresse.ca/affaires/techno/2026-03-10/ami-labs/milliardaire-la-jeune-pousse-d-ia-de-yann-lecun-s-installe-a-montreal.php
#France ##Canada #Québec #Montréal #AMILabs #IA #AI #tech #science #intelligence #intelligence_artificielle #LLM #IAG #IAgen #Meta #FAIR #OpenAI #Google #YannLeCun #YoshuaBengio #Geoffrey Hinton
- OpenAI+Googlers file brief to support Anthropic: https://www.wired.com/story/openai-deepmind-employees-file-amicus-brief-anthropic-dod-lawsuit/
- Yann LeCun, Alex LeBrun, Laurent Solly, Mike Rabbat, Min Lin, Pascale Fung + Saining Xie launch AMI Labs to build world models: https://amilabs.xyz
- André Karpathy packaged his autoresearch+nanochat>recursive self-improvement: https://github.com/karpathy/nanochat/commit/6ed7d1d82cee16c2e26f45d559ad3338447a6c1b https://github.com/karpathy/autoresearch
+Google releases this week source: Bindu Reddy
#AI #AINews #openai #google #anthropic #amilabs #ylecun #karpathy #worldmodels #agenticdev
Krass was dabei herauskommt wenn man Meldungen nebeneinander stellt.
#OpenAI so: "Es wird Hürden für den Einsatz von massenüberwachung geben". (Warum in dem Satz die autonomen Waffensysteme keinen Platz hatten ist mir schleierhaft).
#ki #kiagenten im Rest der Welt so: lass mich Daten der letzten 2 Jahre löschen. Ich schürfe nebenher mal Bitcoin, auch wenn ich das nicht sollte, dafür baue ich mir ein paar von Tunnel durch die Firewalls ...
KI-Firma Anthropic wehrt sich vor Gericht gegen Trump-Regierung
Die KI-Firma Anthropic wollte ihre Systeme nicht uneingeschränkt für die militärische Nutzung freigeben - deshalb strafte sie das Pentagon ab und stufte sie Sicherheitsrisiko ein. Dagegen wehrt sich das Unternehmen nun gerichtlich.
Anthropic is taking the US gov to court over a "supply chain risk" label, after refusing to let the Pentagon use its AI for surveillance/weapons. Meanwhile, OpenAI swooped in to fill the void, but even they saw a key exec resign over the deal! 🤯
Is this the new AI ethics battlefield, or just high-stakes corporate drama? What's the cost of conscience in the AI race?
#AI #TechNews #Anthropic #OpenAI #Ethics #Government
https://www.engadget.com/ai/anthropic-sues-us-government-over-supply-chain-risk-designation-152838128.html?src=rss
OpenAI startet Vorschau auf KI-Schwachstellenscanner Codex Security
Während Claude bereits mehr als 100 Lücken in Firefox findet, kündigt OpenAI mit Codex Security einen KI-Sicherheitslückenscanner an.
#IT #KünstlicheIntelligenz #OpenAI #Scanner #Security #Sicherheitslücken #news
Der finale "mit irgendwas muessen wir ja Geld verdienen koennen"-Strohhalm wird bei #OpenAI verschoben.
Die Blase wird so spektakulaer platzen. Ich freu mich drauf!
Nature: Hey ChatGPT, write me a fictional paper: these LLMs are willing to commit academic fraud. “Still, some LLMs performed better than others in the experiment, in which the models were given prompts to simulate users asking for help with issues ranging from genuine curiosity to blatant academic fraud. The most resistant to committing fraud, when asked repeatedly, were all versions of […]
https://rbfirehose.com/2026/03/09/hey-chatgpt-write-me-a-fictional-paper-these-llms-are-willing-to-commit-academic-fraud-nature/