• chevron_right

      “The king is dead”—Claude 3 surpasses GPT-4 on Chatbot Arena for the first time

      news.movim.eu / ArsTechnica · 2 days ago - 16:32 · 1 minute

    Two toy robots fighting, one knocking the other's head off.

    Enlarge (credit: Getty Images / Benj Edwards )

    On Tuesday, Anthropic's Claude 3 Opus large language model (LLM) surpassed OpenAI's GPT-4 (which powers ChatGPT) for the first time on Chatbot Arena , a popular crowdsourced leaderboard used by AI researchers to gauge the relative capabilities of AI language models. "The king is dead," tweeted software developer Nick Dobos in a post comparing GPT-4 Turbo and Claude 3 Opus that has been making the rounds on social media. "RIP GPT-4."

    Since GPT-4 was included in Chatbot Arena around May 10, 2023 (the leaderboard launched May 3 of that year), variations of GPT-4 have consistently been on the top of the chart until now, so its defeat in the Arena is a notable moment in the relatively short history of AI language models. One of Anthropic's smaller models, Haiku, has also been turning heads with its performance on the leaderboard.

    "For the first time, the best available models—Opus for advanced tasks, Haiku for cost and efficiency—are from a vendor that isn't OpenAI," independent AI researcher Simon Willison told Ars Technica. "That's reassuring—we all benefit from a diversity of top vendors in this space. But GPT-4 is over a year old at this point, and it took that year for anyone else to catch up."

    Read 8 remaining paragraphs | Comments

    • chevron_right

      GPT-3.5 champion de Street Fighter III

      news.movim.eu / Korben · 3 days ago - 14:32 · 3 minutes

    J’espère que vous êtes en forme et prêts à en découdre, car aujourd’hui on va parler d’un sujet marrant : GPT-3.5 Turbo d’OpenAI est devenu le nouveau champion toutes catégories de Street Fighter III !

    Non, j’ai rien fumé, il y a bien une IA qui a mis la pâtée à tous ses adversaires lors d’un tournoi un peu spécial.

    En effet, la semaine dernière, lors du Mistral AI Hackathon à San Francisco, une équipe de passionnés a eu l’idée de génie d’organiser un tournoi un peu particulier. : Faire s’affronter différents modèles de langage sur le cultissime jeu de baston Street Fighter III , pour voir lequel allait sortir vainqueur.

    Parce que bon, c’est bien beau de savoir faire la conversation ou générer des images moches, mais quand il s’agit d’envoyer des tatanes dans la tronche, il faut être un peu plus réactif !

    Et c’est là que notre pote GPT-3.5 sort les muscles et s’en sort très bien. Contrairement aux algorithmes d’apprentissage par renforcement (deep learning) qui se contentent bêtement d’accumuler des points en fonction de leurs actions, les modèles de langage comme GPT sont capables de comprendre un contexte et d’agir en conséquence.

    En gros, ils analysent ce qu’il se passe à l’écran, les mouvements des personnages, leur barre de vie… Et en fonction de ça, ils décident quelle attaque lancer. Un peu comme un joueur humain en fait, sauf qu’eux n’ont pas besoin de café pour rester concentrés.

    Les premières bagarres ont opposé différentes versions du modèle Mistral , dans des combats endiablés dignes des plus grands shōnens. Mais très vite, l’équipe a décidé de corser un peu les choses en invitant OpenAI et ses modèles GPT-3.5 et GPT-4 dans l’arène. Et là, mes amis, ça a commencé à sentir le roussi pour la concurrence !

    Les poings ont volé, les combos se sont enchaînés, les contres se sont succédés à un rythme infernal. Un vrai feu d’artifice d’uppercuts, de coups spéciaux et de provocations bien senties. Mais au final, après des dizaines de combats acharnés, c’est bien GPT-3.5 (et plus précisément sa dernière version « Turbo ») qui est ressorti vainqueur ! La médaille d’argent revient à Mistral-small-2042, qui a réussi l’exploit de coiffer sur le poteau un modèle GPT-4 en accès anticipé.

    Tout ça pour dire que si vous voulez vous mesurer à ces champions, c’est tout à fait possible ! Le code source du projet est disponible sur Github , et vous n’avez même pas besoin d’un supercalculateur pour faire tourner tout ça. Il vous faudra juste dénicher une ROM de jeu de baston 2D ou 3D old school, et le tour est joué. Perso j’ai hâte de voir ce que ça donne sur un bon vieux Tekken 3…

    Pour installer et tester LLM Colosseum :

    1. Suivez les instructions de la documentation DIAMBRA , l’outil qui permet de faire jouer les LLM
    2. Téléchargez la ROM et placez-la dans ~/.diambra/roms
    3. Clonez le dépôt de llm coloseum et installez les paquets Python requis avec la commande pip3 install -r requirements.txt
    4. Créez un fichier nommé .env et copiez-y le contenu du fichier .env.example
    5. Lancez le programme avec la commande make run

    Blague à part, cette expérience montre bien le potentiel hallucinant des modèles de langage pour les jeux vidéo. On peut tout à fait imaginer des PNJ avec lesquels on pourrait interagir de façon totalement naturelle et immersive, des adversaires capables de s’adapter à votre style de jeu et de vous surprendre… Bref, de quoi révolutionner complètement notre façon de jouer ! Après, faudra quand même faire gaffe à pas trop les énerver, on a bien vu ce que ça donnait quand on laissait GPT-3.5 jouer à des wargames … Boum, plus de planète !

    Allez, je vous laisse, faut que je retourne taper Zangief moi.

    Merci à Lorenper pour l’info et à très vite pour de nouvelles aventures.

    • chevron_right

      Vernor Vinge, father of the tech singularity, has died at age 79

      news.movim.eu / ArsTechnica · Thursday, 21 March - 15:33

    A photo of Vernor Vinge in 2006. https://en.wikipedia.org/wiki/Vernor_Vinge#/media/File:Vernor_Vinge.jpg

    Enlarge / A photo of Vernor Vinge in 2006. (credit: Raul654 )

    On Wednesday, author David Brin announced that Vernor Vinge, sci-fi author, former professor, and father of the technological singularity concept, died from Parkinson's disease at age 79 on March 20, 2024, in La Jolla, California. The announcement came in a Facebook tribute where Brin wrote about Vinge's deep love for science and writing.

    "A titan in the literary genre that explores a limitless range of potential destinies, Vernor enthralled millions with tales of plausible tomorrows, made all the more vivid by his polymath masteries of language, drama, characters, and the implications of science," wrote Brin in his post.

    As a sci-fi author, Vinge won Hugo Awards for his novels A Fire Upon the Deep (1993), A Deepness in the Sky (2000), and Rainbows End (2007). He also won Hugos for novellas Fast Times at Fairmont High (2002) and The Cookie Monster (2004). As Mike Glyer's File 770 blog notes, Vinge's novella True Names (1981) is frequency cited as the first presentation of an in-depth look at the concept of "cyberspace."

    Read 8 remaining paragraphs | Comments

    • chevron_right

      GPT-5 : le prochain modèle d’OpenAI n’est plus très loin

      news.movim.eu / JournalDuGeek · Thursday, 21 March - 14:38

    OpenAI DevDay

    Des sources anonymes affirment que la prochaine version du LLM s'annonce très impressionnante, et qu'elle pourrait débarquer cet été.
    • chevron_right

      GPT-5 might arrive this summer as a “materially better” update to ChatGPT

      news.movim.eu / ArsTechnica · Wednesday, 20 March - 21:53 · 1 minute

    A glowing OpenAI logo on a blue background.

    Enlarge

    When OpenAI launched its GPT-4 AI model a year ago, it created a wave of immense hype and existential panic from its ability to imitate human communication and composition. Since then, the biggest question in AI has remained the same: When is GPT-5 coming out? During interviews and media appearances around the world, OpenAI CEO Sam Altman frequently gets asked this question, and he usually gives a coy or evasive answer , sometimes coupled with promises of amazing things to come.

    According to a new report from Business Insider , OpenAI is expected to release GPT-5, an improved version of the AI language model that powers ChatGPT, sometime in mid-2024—and likely during the summer. Two anonymous sources familiar with the company have revealed that some enterprise customers have recently received demos of GPT-5 and related enhancements to ChatGPT .

    One CEO who recently saw a version of GPT-5 described it as "really good" and "materially better," with OpenAI demonstrating the new model using use cases and data unique to his company. The CEO also hinted at other unreleased capabilities of the model, such as the ability to launch AI agents being developed by OpenAI to perform tasks automatically.

    Read 7 remaining paragraphs | Comments

    • chevron_right

      Mustafa Suleyman: the new head of Microsoft AI with concerns about his trade

      news.movim.eu / TheGuardian · Wednesday, 20 March - 19:19

    The son of a Syrian taxi driver and an English nurse torn between AI’s potential and problems

    Like many artificial intelligence pioneers, Mustafa Suleyman has expressed concerns about a technology he has played an important role in developing.

    Speaking at the global AI safety summit last year , the 39-year-old Briton said there might have to be a pause in development towards the end of the decade. “I don’t rule it out. And I think that at some point over the next five years or so, we’re going to have to consider that question very seriously,” he said.

    Continue reading...
    • chevron_right

      Nvidia unveils Blackwell B200, the “world’s most powerful chip” designed for AI

      news.movim.eu / ArsTechnica · Tuesday, 19 March - 15:27 · 1 minute

    The GB200 "superchip" covered with a fanciful blue explosion that suggests computational power bursting forth from within. The chip does not actually glow blue in reality.

    Enlarge / The GB200 "superchip" covered with a fanciful blue explosion that suggests computational power bursting forth from within. The chip does not actually glow blue in reality. (credit: Nvidia / Benj Edwards)

    On Monday, Nvidia unveiled the Blackwell B200 tensor core chip—the company's most powerful single-chip GPU, with 208 billion transistors—which Nvidia claims can reduce AI inference operating costs (such as running ChatGPT ) and energy consumption by up to 25 times compared to the H100 . The company also unveiled the GB200, a "superchip" that combines two B200 chips and a Grace CPU for even more performance.

    The news came as part of Nvidia's annual GTC conference, which is taking place this week at the San Jose Convention Center. Nvidia CEO Jensen Huang delivered the keynote Monday afternoon. "We need bigger GPUs," Huang said during his keynote. The Blackwell platform will allow the training of trillion-parameter AI models that will make today's generative AI models look rudimentary in comparison, he said. For reference, OpenAI's GPT-3, launched in 2020, included 175 billion parameters. Parameter count is a rough indicator of AI model complexity.

    Nvidia named the Blackwell architecture after David Harold Blackwell , a mathematician who specialized in game theory and statistics and was the first Black scholar inducted into the National Academy of Sciences. The platform introduces six technologies for accelerated computing, including a second-generation Transformer Engine, fifth-generation NVLink, RAS Engine, secure AI capabilities, and a decompression engine for accelerated database queries.

    Read 8 remaining paragraphs | Comments

    • chevron_right

      New York Times: Microsoft’s AI Tools Are Nothing Like The VCR

      news.movim.eu / TorrentFreak · Tuesday, 19 March - 15:10 · 4 minutes

    betamax “The VCR is to the American film producer and the American public as the Boston Strangler is to the woman home alone.”

    The quote above comes from the late Jack Valenti , who was the Motion Picture Association’s boss in 1982, when he warned the House of Representatives of the looming video recorder threat.

    With the benefit of hindsight, the VCR wasn’t all that scary for Hollywood. The movie industry continued to flourish in the decades that followed, while technology continued along the path of progress too. New inventions have come along and for many rightsholders, generative AI (GenAI) is today’s growing concern.

    The VCR threat cemented itself in legal history through the Betamax decision which still plays a role today, for various reasons. As reported yesterday , the decision was cited in the ongoing legal battle between book publishers and the Internet Archive. Simultaneously, the VCR is also starring in a legal dispute between the New York Times and Microsoft.

    NYT vs. OpenAI/Microsoft

    Late last year The Times sued OpenAI and Microsoft , alleging that its generative AI models were trained on copyrighted news articles. The news publication also suggested that, when prompted the right way, ChatGPT could recite content from these articles.

    OpenAI previously asked the court to dismiss these claims , alleging that the New York Times ‘hacked’ its service to produce the ‘highly anomalous’ outputs. The Times’ DMCA violation claim, misappropriation claim, and contributory infringement claim either fail or fall short, OpenAI added.

    Microsoft filed a separate motion to dismiss. Instead of focusing on alleged hacking practices, the company led with Jack Valenti’s ‘Boston Strangler’ analogy, part of an all-out effort by television and movie producers to stop a groundbreaking new technology.

    Hollywood was ultimately unable to stop the VCR. The ‘scaremongering’ didn’t convince the Supreme Court in ‘Sony (Betamax) v. Universal City Studios’ and the VCR was declared legal. The rest is history.

    Microsoft Cites VCR Scaremongering

    According to Microsoft, the legal crusade against AI models should be seen similarly. Instead of alleging concrete copyright infringement by end users, the technology itself is framed as copyright-infringing. That’s incorrect, the tech giant countered in its motion to dismiss earlier this month.

    “At most, The Times’s allegations establish Microsoft’s awareness that someone could use a GPT-based product to infringe. Of course, the same was true of the VCR — as it is of word processors, hard drives, social media feeds, internet connections, and so forth.

    “Fortunately, the Supreme Court long ago rejected liability merely based on offering a multi-use product,” Microsoft added.

    Microsoft asked the court to dismiss several key claims, including contributory copyright infringement. The claim fails because there’s no evidence that the tech company knew of any third-party copyright infringements or contributed to them, Microsoft argued.

    NYT: AI is Nothing Like the VCR

    Yesterday, The Times responded to the motion to dismiss, starting with a VCR analogy. According to the news outlet, the VCR is nothing like the GenAI threat they’re facing today.

    “Defendants’ generative AI models are nothing like VCRs. Sony didn’t copy movies and television shows to build VCRs; Defendants built their AI models by copying millions of Times articles and other copyrighted works without permission or payment.”

    Unlike the VCR, GPT services are trained on copyrighted content without permission and can reproduce these in part, The Times argues.

    “Defendants are using their AI models to copy and summarize even breaking news articles that users would otherwise seek on a publisher’s website. If VCRs had been built with movies to make movies that compete with movies, or if Sony oversaw the VCR’s infringing users, Sony would have gone the other way.”

    Microsoft’s defense is largely predicated on its conclusion that the use of copyrighted content for AI training is fair use. The Times sees this as an “absurd argument” but doesn’t respond to it in detail. Instead, it mostly sticks to its original claims.

    ‘Using AI to Bypass Paywalls’

    One of the key claims in the complaint is contributory copyright infringement. According to legal precedents, a party can be found liable for copyright infringement if it induces, causes, or materially contributes to it. This is particularly true when a service has few non-infringing uses.

    This claim was also used against Sony’s VCR but that ultimately failed. The Times hasn’t listed any concrete infringements in its complaint but notes that its pleadings against Microsoft are sufficient to survive a motion to dismiss at this stage.

    “Although Microsoft argues that copyright infringement by users of its GenAI products is just a ‘theoretical possibility’, the question at this stage is whether The Times has plausibly alleged that such infringement has taken place. The answer is yes,” Microsoft notes.

    To illustrate, The Times references a Gizmodo article that suggested ChatGPT’s ‘Browse with Bing’ was paused after people used it to bypass paywalls. Microsoft wasn’t blind to these copyright infringement issues, The Times notes, stressing that it previously alerted the company to its concerns.

    All in all, the Times wants the case to move forward in its entirety while Microsoft would like it to end here. It’s now up to the court to decide if the case can go forward, and on what claims. Alternatively, the parties can choose to settle their disagreements outside of court but, thus far, there’s no evidence to show either side prefers that option.

    —-

    A copy of Microsoft’s motion to dismiss, submitted at a New York federal court, can be found here (pdf) and The Times’ response is available here (pdf)

    From: TF , for the latest news on copyright battles, piracy and more.