ChatGPT vs Google Bard: Which is better? We put them to the test.

We compare two top AI language models in seven categories to pick a winner.

An AI-generated image of two robots fighting in an arena.

Enlarge / An AI-generated image of two robots fighting in an arena. (credit: Benj Edwards / Midjourney)

In today's world of generative AI chatbots, we've witnessed the sudden rise of OpenAI's ChatGPT, introduced in November, followed by Bing Chat in February and Google's Bard in March. We decided to put these chatbots through their paces with an assortment of tasks to determine which one reigns supreme in the AI chatbot arena. Since Bing Chat uses similar GPT-4 technology as the latest ChatGPT model, we opted to focus on two titans of AI chatbot technology: OpenAI and Google.

We tested ChatGPT and Bard in seven critical categories: dad jokes, argument dialog, mathematical word problems, summarization, factual retrieval, creative writing, and coding. For each test, we fed the exact same instruction (called a "prompt") into ChatGPT (with GPT-4) and Google Bard. We used the first result, with no cherry-picking.

It's worth noting that a version of ChatGPT based on the earlier GPT-3.5 model is also available, but we did not use that in the test. Since we used GPT-4 only, we will refer to ChatGPT as "ChatGPT-4" in this article to reduce confusion.

Read 45 remaining paragraphs | Comments

ChatGPT vs Google Bard: Which is better? We put them to the test.

We compare two top AI language models in seven categories to pick a winner.

An AI-generated image of two robots fighting in an arena.

Enlarge / An AI-generated image of two robots fighting in an arena. (credit: Benj Edwards / Midjourney)

In today's world of generative AI chatbots, we've witnessed the sudden rise of OpenAI's ChatGPT, introduced in November, followed by Bing Chat in February and Google's Bard in March. We decided to put these chatbots through their paces with an assortment of tasks to determine which one reigns supreme in the AI chatbot arena. Since Bing Chat uses similar GPT-4 technology as the latest ChatGPT model, we opted to focus on two titans of AI chatbot technology: OpenAI and Google.

We tested ChatGPT and Bard in seven critical categories: dad jokes, argument dialog, mathematical word problems, summarization, factual retrieval, creative writing, and coding. For each test, we fed the exact same instruction (called a "prompt") into ChatGPT (with GPT-4) and Google Bard. We used the first result, with no cherry-picking.

It's worth noting that a version of ChatGPT based on the earlier GPT-3.5 model is also available, but we did not use that in the test. Since we used GPT-4 only, we will refer to ChatGPT as "ChatGPT-4" in this article to reduce confusion.

Read 45 remaining paragraphs | Comments

Super Mario Bros Movie: Das Schnurrbart-gegen-Schildkröte-Spektakel

Früher der kleine Game Boy, jetzt die große Kinoleinwand – kann das gutgehen? Wir sind trotz der holzschnittartigen Handlung von Super Mario Bros Movie begeistert! Eine Rezension von Peter Steinlechner (Super Mario, Nintendo)

Früher der kleine Game Boy, jetzt die große Kinoleinwand - kann das gutgehen? Wir sind trotz der holzschnittartigen Handlung von Super Mario Bros Movie begeistert! Eine Rezension von Peter Steinlechner (Super Mario, Nintendo)

(g+) Sonarqube für Codequalität: Besser als furchtbar ist noch lange nicht gut

Im schnellen Alltag bleibt die Qualität des Codes oft auf der Strecke. Wir zeigen, wie sie mit Sonarqube und CI/CD einfach gemessen und auf Einhaltung von Standards überprüft werden kann. Eine Anleitung von Rene Koch (Softwareentwicklung, Software)

Im schnellen Alltag bleibt die Qualität des Codes oft auf der Strecke. Wir zeigen, wie sie mit Sonarqube und CI/CD einfach gemessen und auf Einhaltung von Standards überprüft werden kann. Eine Anleitung von Rene Koch (Softwareentwicklung, Software)

Take-Two Dismisses Claims Against Lead Defendants in GTA Mods Lawsuit

In 2021, Take-Two Interactive sued several people said to be behind the popular re3 and reVC Grand Theft Auto fan projects. Four key named defendants fired back, arguing that their game fixes & enhancements were fair use, not piracy. Despite having identified no other defendants, Take-Two has just dismissed its case against the four men, with prejudice.

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

gtaIn 2021, a group of Grand Theft Auto enthusiast programmers released ‘re3’ and ‘reVC’, a pair of reverse-engineered modifications for GTA 3 and Vice City.

These projects breathed new life into games that while still fantastic, benefited greatly from significant enhancements that helped to wind back the years. Fans loved ‘re3’ and ‘reVC’ but Take-Two and Rockstar Games most definitely did not.

The companies’ first move was to file a DMCA notice that ordered GitHub to take the projects down. The programmers’ response came via a DMCA counternotice which restored their projects on GitHub but left them exposed to a potential legal response by Take-Two.

Take-Two Files Copyright Lawsuit

In September 2021, Take-Two filed a lawsuit against the programmers, claiming that the aim of the projects was to create and distribute pirated versions of GTA 3 and Vice City.

The company’s claims included damages for “willful and malicious” copyright infringement due to the illegal copying, adaption and distribution of GTA source code and other protected content. For good measure, Take-Two also demanded damages for alleged misrepresentations in the defendants’ DMCA counternotices.

The lawsuit listed 14 defendants, just four of which were named: Angelo Papenhoff (aap), Theo Morra, Eray Orçunus, and Adrian Graber. In their November 2021 answer to the complaint, the four men denied the copyright infringement allegations and cited fair use among other affirmative defenses.

One Year Later, Limited Visible Progress

The first five months of 2022 were relatively uneventful, at least based on information made available to the public. The parties did participate in an ADR (Alternative Dispute Resolution) session but according to the mediator’s report early May, the case did not settle.

No further entries appeared on the docket until December 2022, when the parties informed the court that since disclosure and discovery in the lawsuit was likely to involve the production of confidential, proprietary, or private information, special protection from public disclosure may be warranted. The court had no problem with granting the request.

As part of this process, Take-Two served an expert report on November 17, 2022, leaving a deadline of December 15 for Papenhoff, Morra, Orçunus, and Graber to disclose their rebuttal expert witness. What followed was a series of requests for time extensions and a note to the court on February 6, 2023, indicating that a settlement had been agreed in principle.

Take-Two Dismisses Claims Against Four Named Defendants

Following the news of the potential settlement, the court granted a 30-day extension until March 19. When that date disappeared into history with no further docket activity, the court issued an order for the parties to file a status report by April 3. The parties responded on the day of the deadline with a joint stipulation of dismissal.

“In accordance with Federal Rule of Civil Procedure 41(a)(1)(A)(ii), by and between the
undersigned counsel for the Plaintiff, Take Two Interactive Software, Inc., and counsel for Defendants Angelo Papenhoff, Theo Morra, Eray Orçunus, and Adrian Graber that all claims asserted in the above-referenced action against the Named Defendants are dismissed with prejudice,” the parties informed the court.

In these circumstances, the dismissal does not require the court to issue an order, since the signatures of the parties who have appeared in the case will suffice. Take-Two and the developers will bear their own costs, expenses, and attorneys’ fees.

What prompted this meeting of minds and subsequent agreement isn’t mentioned, but for the four men, the lawsuit is over and cannot be refiled in the future.

On Monday, the lawsuit appeared set to continue against the remaining defendants – Doe 1 (a/k/a Ash R. and ASH_735) plus Does 2 through 10 inclusive – whoever (and wherever) they might be. An update filed Tuesday clarified the position.

“Plaintiff, Take Two Interactive Software, Inc., hereby voluntarily dismisses Defendants Doe 1 a/k/a ASH R. and ASH_735 and Does 2-10, without prejudice,” the company informed the court.

It’s possible that a fresh lawsuit will be filed in the future but, realistically, not anytime soon.

The joint stipulation of dismissal can be found here (pdf)

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