Was OpenAI GPT-4o Hype A Troll On Google? via @sejournal, @martinibuster

OpenAI managed to steal the attention away from Google in the weeks leading up to Google’s biggest event of the year (Google I/O). When the big announcement arrived there all they had to show was a language model that was slightly better than the previous one with the “magic” part not even in Alpha testing stage.

OpenAI may have left users feeling like a mom receiving a vacuum cleaner for Mothers Day but it surely succeeded in minimizing press attention for Google’s important event.

The Letter O

The first hint that there’s at least a little trolling going on is the name of the new GPT model, 4 “o” with the letter “o” as in the name of Google’s event,  I/O.

OpenAI says that the letter O stands for Omni, which means everything, but it sure seems like there’s a subtext to that choice.

GPT-4o Oversold As Magic

Sam Altman in a tweet the Friday before the announcement promised “new stuff” that felt like “magic” to him:

“not gpt-5, not a search engine, but we’ve been hard at work on some new stuff we think people will love! feels like magic to me.”

OpenAI co-founder Greg Brockman tweeted:

“Introducing GPT-4o, our new model which can reason across text, audio, and video in real time.

It’s extremely versatile, fun to play with, and is a step towards a much more natural form of human-computer interaction (and even human-computer-computer interaction):”

The announcement itself explained that previous versions of ChatGPT used three models to process audio input. One model to turn audio input into text. Another model to complete the task and output the text version of it and a third model to turn the text output into audio. The breakthrough for GPT-4o is that it can now process the audio input and output within a single model and output it all in the same amount of time that it takes a human to listen and respond to a question.

But the problem is that the audio part isn’t online yet. They’re still working on getting the guardrails working and it will take weeks before an Alpha version is released to a few users for testing. Alpha versions are expected to possibly have bugs while the Beta versions are generally closer to the final products.

This is how OpenAI explained the disappointing delay:

“We recognize that GPT-4o’s audio modalities present a variety of novel risks. Today we are publicly releasing text and image inputs and text outputs. Over the upcoming weeks and months, we’ll be working on the technical infrastructure, usability via post-training, and safety necessary to release the other modalities.

The most important part of GPT-4o, the audio input and output, is finished but the safety level is not yet ready for public release.

Some Users Disappointed

It’s inevitable that an incomplete and oversold product would generate some negative sentiment on social media.

AI engineer Maziyar Panahi (LinkedIn profile) tweeted his disappointment:

“I’ve been testing the new GPT-4o (Omni) in ChatGPT. I am not impressed! Not even a little! Faster, cheaper, multimodal, these are not for me.
Code interpreter, that’s all I care and it’s as lazy as it was before!”

He followed up with:

“I understand for startups and businesses the cheaper, faster, audio, etc. are very attractive. But I only use the Chat, and in there it feels pretty much the same. At least for Data Analytics assistant.

Also, I don’t believe I get anything more for my $20. Not today!”

There are others across Facebook and X that expressed similar sentiments although many others were happy with what they felt was an improvement in speed and cost for the API usage.

Did OpenAI Oversell GPT-4o?

Given that the GPT-4o is in an unfinished state it’s hard not to miss the impression that the release was timed to coincide with and detract from Google I/O. Releasing it on the eve of Google’s big day with a half-finished product may have inadvertently created the impression that GPT-4o in the current state is a minor iterative improvement.

In the current state it’s not a revolutionary step forward but once the audio portion of the model exits Alpha testing stage and makes it through the Beta testing stage then we can start talking about revolutions in large language model. But by the time that happens Google and Anthropic may already have staked a flag on that mountain.

OpenAI’s announcement paints a lackluster image of the new model, promoting the performance as on the same level as GPT-4 Turbo. The only bright spots is the significant improvements in languages other than English and for API users.

OpenAI explains:

• “It matches GPT-4 Turbo performance on text in English and code, with significant improvement on text in non-English languages, while also being much faster and 50% cheaper in the API.”

Here are the ratings across six benchmarks that shows GPT-4o barely squeaking past GPT-4T in most tests but falling behind GPT-4T in an important benchmark for reading comprehension.

Here are the scores:

• MMLU (Massive Multitask Language Understanding)
  This is a benchmark for multitasking accuracy and problem solving in over fifty topics like math, science, history and law. GPT-4o (scoring 88.7) is slightly ahead of GPT4 Turbo (86.9).
• GPQA (Graduate-Level Google-Proof Q&A Benchmark)
  This is 448 multiple-choice questions written by human experts in various fields like biology, chemistry, and physics. GPT-4o scored 53.6, slightly outscoring GPT-4T (48.0).
• Math
  GPT 4o (76.6) outscores GPT-4T by four points (72.6).
• HumanEval
  This is the coding benchmark. GPT-4o (90.2) slightly outperforms GPT-4T (87.1) by about three points.
• MGSM (Multilingual Grade School Math Benchmark)
  This tests LLM grade-school level math skills across ten different languages. GPT-4o scores 90.5 versus 88.5 for GPT-4T.
• DROP (Discrete Reasoning Over Paragraphs)
  This is a benchmark comprised of 96k questions that tests language model comprehension over the contents of paragraphs. GPT-4o (83.4) scores nearly three points lower than GPT-4T (86.0).

Did OpenAI Troll Google With GPT-4o?

Given the provocatively named model with the letter o, it’s hard to not consider that OpenAI is trying to steal media attention in the lead-up to Google’s important I/O conference. Whether that was the intention or not OpenAI wildly succeeded in minimizing attention given to Google’s upcoming search conference.

Does a language model that barely outperforms its predecessor worth all the hype and media attention it received? The pending announcement dominated news coverage over Google’s big event so for OpenAI the answer is clearly yes, it was worth the hype.

Featured Image by Shutterstock/BeataGFX

OpenAI Announces ChatGPT 4o Omni via @sejournal, @martinibuster

ChatGPT announced a new version of ChatGPT that can accept audio, image and text inputs and also generate outputs in audio, image and text. OpenAI is calling the new version of ChatGPT 4o, with the “o” standing for “omni” which is a combining form word that means “all”.

ChatGPT 4o (Omni)

OpenAI described this new version of ChatGPT as a progression toward more natural human and machine interactions which responds to user inputs at the same speed as a human to human conversations. The new version matches ChatGPT 4 Turbo in English and significantly outperforms Turbo in other languages. There is a significant improvement in API performance, increasing in speed and operating 50% less expensively.

The announcement explains:

“As measured on traditional benchmarks, GPT-4o achieves GPT-4 Turbo-level performance on text, reasoning, and coding intelligence, while setting new high watermarks on multilingual, audio, and vision capabilities.”

Advanced Voice Processing

The previous method for communicating with voice involved bridging together three different models to handle transcribing voice inputs to text where the second model (GPT 3.5 or GPT-4) processes it and outputs text and a third model that transcribes the text back into audio. That method is said to lose nuances in the various translations.

OpenAI described the downsides of the previous approach that are (presumably) overcome by the new approach:

“This process means that the main source of intelligence, GPT-4, loses a lot of information—it can’t directly observe tone, multiple speakers, or background noises, and it can’t output laughter, singing, or express emotion.”

The new version doesn’t need three different models because all of the inputs and outputs are handled together in one model for end to end audio input and output. Interestingly, OpenAI states that they haven’t yet explored the full capabilities of the new model or fully understand the limitations of it.

New Guardrails And An Iterative Release

OpenAI GPT 4o features new guardrails and filters to keep it safe and avoid unintended voice outputs for safety. However today’s announcement says that they are only rolling out the capabilities for text and image inputs and text outputs and a limited audio at launch. GPT 4o is available for both free and paid tiers, with Plus users receiving 5 times higher message limits.

Audio capabilities are due for a limited alpha-phase release for ChatGPT Plus and API users within weeks.

The announcement explained:

“We recognize that GPT-4o’s audio modalities present a variety of novel risks. Today we are publicly releasing text and image inputs and text outputs. Over the upcoming weeks and months, we’ll be working on the technical infrastructure, usability via post-training, and safety necessary to release the other modalities. For example, at launch, audio outputs will be limited to a selection of preset voices and will abide by our existing safety policies.”

Read the announcement:

Hello GPT-4o

Featured Image by Shutterstock/Photo For Everything

OpenAI Expected to Integrate Real-Time Data In ChatGPT via @sejournal, @martinibuster

Sam Altman, CEO of OpenAI, dispelled rumors that a new search engine would be announced on Monday, May 13. Recent deals have raised the expectation that OpenAI will announce the integration of real-time content from English, Spanish, and French publications into ChatGPT, complete with links to the original sources.

OpenAI Search Is Not Happening

Many competing search engines have tried and failed to challenge Google as the leading search engine. A new wave of hybrid generative AI search engines is currently trying to knock Google from the top spot with arguably very little success.

Sam Altman is on record saying that creating a search engine to compete against Google is not a viable approach. He suggested that technological disruption was the way to replace Google by changing the search paradigm altogether. The speculation that Altman is going to announce a me-too search engine on Monday never made sense given his recent history of dismissing the concept as a non-starter.

So perhaps it’s not a surprise that he recently ended the speculation by explicitly saying that he will not be announcing a search engine on Monday.

He tweeted:

“not gpt-5, not a search engine, but we’ve been hard at work on some new stuff we think people will love! feels like magic to me.”

“New Stuff” May Be Iterative Improvement

It’s quite likely that what’s going to be announced is iterative which means it improves ChatGPT but not replaces it. This fits into how Altman recently expressed his approach with ChatGPT.

He remarked:

“And it does kind of suck to ship a product that you’re embarrassed about, but it’s much better than the alternative. And in this case in particular, where I think we really owe it to society to deploy iteratively.

There could totally be things in the future that would change where we think iterative deployment isn’t such a good strategy, but it does feel like the current best approach that we have and I think we’ve gained a lot from from doing this and… hopefully the larger world has gained something too.”

Improving ChatGPT iteratively is Sam Altman’s preference and recent clues point to what those changes may be.

Recent Deals Contain Clues

OpenAI has been making deals with news media and User Generated Content publishers since December 2023. Mainstream media has reported these deals as being about licensing content for training large language models. But they overlooked a a key detail that we reported on last month which is that these deals give OpenAI access to real-time information that they stated will be used to give attribution to that real-time data in the form of links.

That means that ChatGPT users will gain the ability to access real-time news and to use that information creatively within ChatGPT.

Dotdash Meredith Deal

Dotdash Meredith (DDM) is the publisher of big brand publications such as Better Homes & Gardens, FOOD & WINE, InStyle, Investopedia, and People magazine. The deal that was announced goes way beyond using the content as training data. The deal is explicitly about surfacing the Dotdash Meredith content itself in ChatGPT.

The announcement stated:

“As part of the agreement, OpenAI will display content and links attributed to DDM in relevant ChatGPT responses. …This deal is a testament to the great work OpenAI is doing on both fronts to partner with creators and publishers and ensure a healthy Internet for the future.

Over 200 million Americans each month trust our content to help them make decisions, solve problems, find inspiration, and live fuller lives. This partnership delivers the best, most relevant content right to the heart of ChatGPT.”

A statement from OpenAI gives credibility to the speculation that OpenAI intends to directly show licensed third-party content as part of ChatGPT answers.

OpenAI explained:

“We’re thrilled to partner with Dotdash Meredith to bring its trusted brands to ChatGPT and to explore new approaches in advancing the publishing and marketing industries.”

Something that DDM also gets out of this deal is that OpenAI will enhance DDM’s in-house ad targeting in order show more tightly focused contextual advertising.

Le Monde And Prisa Media Deals

In March 2024 OpenAI announced a deal with two global media companies, Le Monde and Prisa Media. Le Monde is a French news publication and Prisa Media is a Spanish language multimedia company. The interesting aspects of these two deals is that it gives OpenAI access to real-time data in French and Spanish.

Prisa Media is a global Spanish language media company based in Madrid, Spain that is comprised of magazines, newspapers, podcasts, radio stations, and television networks. It’s reach extends from Spain to America. American media companies include publications in the United States, Argentina, Bolivia, Chile, Colombia, Costa Rica, Ecuador, Mexico, and Panama. That is a massive amount of real-time information in addition to a massive audience of millions.

OpenAI explicitly announced that the purpose of this deal was to bring this content directly to ChatGPT users.

The announcement explained:

“We are continually making improvements to ChatGPT and are supporting the essential role of the news industry in delivering real-time, authoritative information to users. …Our partnerships will enable ChatGPT users to engage with Le Monde and Prisa Media’s high-quality content on recent events in ChatGPT, and their content will also contribute to the training of our models.”

That deal is not just about training data. It’s about bringing current events data to ChatGPT users.

The announcement elaborated in more detail:

“…our goal is to enable ChatGPT users around the world to connect with the news in new ways that are interactive and insightful.”

As noted in our April 30th article that revealed that OpenAI will show links in ChatGPT, OpenAI intends to show third party content with links to that content.

OpenAI commented on the purpose of the Le Monde and Prisa Media partnership:

“Over the coming months, ChatGPT users will be able to interact with relevant news content from these publishers through select summaries with attribution and enhanced links to the original articles, giving users the ability to access additional information or related articles from their news sites.”

There are additional deals with other groups like The Financial Times which also stress that this deal will result in a new ChatGPT feature that will allow users to interact with real-time news and current events .

OpenAI’s Monday May 13 Announcement

There are many clues that the announcement on Monday will be that ChatGPT users will gain the ability to interact with content about current events.  This fits into the terms of recent deals with news media organizations. There may be other features announced as well but this part is something that there are many clues pointing to.

Watch Altman’s interview at Stanford University

Featured Image by Shutterstock/photosince

LinkedIn Report: AI Skills Now Must-Have For Marketers via @sejournal, @MattGSouthern

A new report by Microsoft and LinkedIn reveals the rapid adoption of AI tools and skills in the marketing industry.

According to the 2024 Work Trend Index Annual Report, which surveyed over 31,000 people across 31 countries, marketing professionals who leverage AI enjoy a competitive advantage.

Employers recognize the efficiency gains AI capabilities provide in marketing roles and increasingly seek applicants with those skills.

Karim R. Lakhani, Chair of the Digital Data Design Institute at Harvard, states in the report:

“Marketers are harnessing the power of AI to work smarter, not just faster. It’s enabling them to focus on higher-value, creative work while automating more routine tasks.”

Here are some highlights from the report illustrating the need to develop an AI skill set to remain competitive.

AI Aptitude: The New Must-Have Skill for Marketers

The survey data reveals a strong preference among business leaders for candidates and employees with AI skills.

A majority, 66%, stated they wouldn’t consider hiring candidates lacking AI proficiency.

Further, 71% expressed a preference for less experienced job seekers with AI skills over more seasoned professionals without that expertise.

This inclination was pronounced in creative fields like marketing and design.

Michael Platt, a neuroscience professor at the Wharton School, states in the report:

“AI is redefining what it means to be a competitive marketer in today’s digital landscape. Professionals who can effectively integrate AI into their work are positioning themselves as invaluable assets to their organizations.”

The report indicates that early-career marketers who develop AI skills could benefit significantly.

77% of leaders reported that employees adept at leveraging AI would be trusted with greater responsibilities earlier in their careers than their peers without AI skills.

The AI Arms Race For Top Marketing Talent

Data from LinkedIn shows that job postings highlighting AI tools and applications have seen a 17% increase in application growth compared to those that don’t mention AI.

Additionally, 54% of early-career employees cited access to AI technologies as a key factor influencing their choice of employer.

Organizations that provide AI training and support for their marketing teams likely have an advantage in attracting top talent.

Why SEJ Cares

The widespread adoption of AI in marketing signifies a shift in the skills and capabilities necessary for succeeding in this rapidly evolving industry.

As AI transforms marketing approaches, professionals who fail to adapt risk being left behind.

The 2024 Work Trend Index Annual Report’s findings are relevant to marketing professionals at all levels. They demonstrate that AI proficiency is necessary for career advancement and job market competitiveness.

Additionally, the report highlights businesses’ role in fostering an AI-driven culture.

Companies investing in AI tools, training, and employee support will be better positioned to attract and retain top talent, drive innovation, and achieve better results.

Read the full report.

How This Can Help You

For marketing professionals to succeed in the AI era, the report suggests:

• Prioritize developing AI skills through courses, workshops, training programs, and collaborating with AI practitioners to gain hands-on experience.
• Embrace experimenting with new AI tools and techniques, integrating them into daily workflows to improve efficiency.
• Share AI knowledge actively with colleagues to foster a culture of knowledge sharing and drive organizational AI adoption.
• Highlight AI capabilities during job searches by demonstrating the successful use of AI to drive results in previous roles.
• Choose employers committed to AI adoption that provide access to cutting-edge AI tools and support ongoing learning.

These recommendations can help you future-proof your career and advance in an increasingly competitive field.

Featured Image: eamesBot/Shutterstock

New Anthropic Claude Team Plan Versus ChatGPT Team via @sejournal, @martinibuster

Anthropic announced a new Team plan that gives businesses the opportunity to access more powerful features with increased management and security controls that make it a strong option for companies to consider.

Generative AI For Teams

Modern SaaS technologies for businesses generally come with a team version that allows collaboration within a company and also gives security and control to management so that proprietary documents don’t accidentally leak to the public.

Open AI launched their ChatGPT Team plan in January 2024, which offered a secured workspace for users within a company at a reasonable monthly and yearly subscription model. Anthropic has finally launched their own Team version of Claude with features that exceed what’s offered by ChatGPT Teams.

Claude Team Compares To ChatGPT Team

Claude is known as a great model for creative purposes and a team version that’s more powerful than the regular paid version makes it even more attractive but the important question is how does it compare to ChatGPT Team?

ChatGPT Team is $5/month cheaper (on the yearly billing plan) than Anthropic’s collaborative plan. Otherwise they are both priced at $30/month per user.

Claude Team is a step up from ChatGPT’s Team in one important way and that’s the context window. A context window is a metric of how much data a model can process at one time. The larger the context window the more data the model can analyze in one batch.

ChatGPT Team offers a context window of 32k but Anthropic Claude Team users enjoy a whopping 200k context window which is about 150,000 words or 500 pages of text.

Team Versus Pro Version

When it comes to collaborating within a company, Claude Team is a better value than the regular Pro version because it provides more usage than the regular Pro plan which means that individual users can do more work than users on the Pro plan. The Team version also offers collaborative features like the ability to create a shared database that can be used for projects. Beyond that it offers administrative tools for managing users and billing.

More features are on the way according to the official announcement:

“In the coming weeks, we will be releasing additional collaboration features, including citations from reliable sources to verify AI-generated claims, integrations with data repositories like codebases or CRMs, and iterating with colleagues on AI-generated documents or projects—all while maintaining the highest standards of security and safety.”

Read the announcement here:

Introducing The Claude Team Plan