Confluence for 2.11.24
Real-time translation is becoming a reality. Microsoft brings Copilot to the masses. Developments in image generation, disclosure, and editing. Thoughts on Gemini and how it compares to GPT-4.
It’s been quite a week in generative AI with the introduction by Google of Gemini and Gemini Advanced. More on that below, as well as other items that have our attention at the intersection of generative AI and communication:
Real-Time Translation Is Becoming a Reality
Microsoft Brings Copilot to the Masses
Developments in Image Generation, Disclosure, and Editing
Thoughts on Gemini and How it Compares to GPT-4
Real-Time Translation Is Becoming a Reality
Yes, it’s a video game. But imagine how this technology will transform corporate communication.
When we first saw how quickly and accurately GPT-4 could translate text, we knew that real-time translation was only a matter of time. And if you play Roblox, that time is now. The game has introduced an innovative AI-driven text chat translation feature, capable of converting text across 15 languages in real-time, allowing players around the world to communicate seamlessly via chat. Not only does it translate between language pairs, it also understands the Roblox-specific terminology and slang that brings nuance to gamer language. You can see a bit of how it works here.
We fully expect real-time translation to be a future standard in corporate communication: in texts, in email, and sooner than you probably think, in real-time video. Not only will a non-native speaker be able to listen to the CEO’s town hall in their own language in real-time, but also in the leader’s voice (something Spotify already does for several of its podcasters — and making this real-time is only a question of computing power and latency).
When it arrives, this tech will be a great asset to leaders and communication professionals, and a nice step forward in creating clarity and inclusion within organizations.
Microsoft Brings Copilot to the Masses
GPT-4 is available to anyone, anywhere, on any device — and Microsoft wants to make sure everyone knows it.
Earlier this week in a blog post, Microsoft’s Yusuf Mehdi announced a series of updates coming to Copilot and linked to a 60-second commercial that will air during this evening’s Super Bowl. The commercial emphatically concludes with the words “Anyone. Anywhere. Any device.” While the capability and user interface updates — especially for image editing (more on that in the item below) — are interesting, we think the bigger takeaway for corporate communication professionals is the emphasis on those last words: anyone, anywhere, any device. With the Copilot web, iOS, and Android applications, anyone in the world can access the world’s leading generative AI model for free, on any device. And Microsoft wants to make sure everyone — including the 100 million people watching the Super Bowl this evening — knows it.
This push from Microsoft comes in the same week that Google made major moves of its own with its updated Gemini app and Gemini Ultra model. We noted then that Gemini gives everyone another option to access leading-edge generative AI technology for everyday use. Again: anyone, anywhere, any device. Why does this matter for corporate communications? The more people use these technologies in their personal lives, the higher their expectations will be at work. They will increasingly expect their employers to provide (or allow) access to the powerful tools they use every day, or at the very least to provide guardrails on what employees can and cannot do. From what we’ve seen, many organizations are in front of this — but many are not. If you’re not doing something already, now is the time. The questions — and pressure — from employees will only increase as tools like Copilot and Gemini put the best models in users’ hands for free, anywhere, on any device.
Developments in Image Generation, Disclosure, and Editing
OpenAI, Meta, Apple, Microsoft, and Google all make moves.
The past two weeks saw a flurry of activity in the image generation space. While the details of the announcements differed, of course, many of them cohered around two primary themes: disclosure and editing.
On the disclosure front, both Meta and OpenAI announced steps to label images as AI-generated, as part of a broader push across the industry:
Meta will begin labeling AI-generated images on Facebook, Instagram, and Threads and is working with companies across the industry to develop standards for watermarking and metadata across different image generation tools. NPR has a nice writeup of Meta’s approach here.
OpenAI, in a similar step, also announced that it would begin watermarking (via metadata) images generated by DALL-E. Given that both ChatGPT and Microsoft Copilot’s image generation capabilities are powered by DALL-E 3, this suggests that a significant amount of all AI-generated images will soon be labeled as such, at least in their metadata. That said, however, as the OpenAI Policy FAQ notes, “Metadata … is not a silver bullet to address issues of provenance. It can easily be removed either accidentally or intentionally.” We agree. These changes are not a silver bullet, and challenges discerning the provenance of content across media are likely to persist, but these actions are steps in the right direction.
The most interesting announcements on the image editing side of things were from Microsoft and Apple:
Microsoft’s Mehdi, in the blog post we referenced earlier about the broader updates to Copilot, notes that with the new updates “you can go beyond just creating images to now customize your generated images with inline editing right inside Copilot, keeping you in the flow of your chat.” He goes on to note that this will include the ability to do things like change colors or backgrounds, resize images, and more without leaving the chat. One of the biggest pain points in image generation to date has been the inability to easily edit the images you create. We do not yet have access to these features in Copilot, but we expect we will soon, and we’re looking forward to experimenting when we do.
While Microsoft’s announcements focused on editing AI-generated images, Apple announced a model that uses natural language to edit existing images. As Emilia David puts it in this helpful writeup in The Verge, this new model “lets users describe in plain language what they want to change in a photo without ever touching photo editing software.” For now, this model is only available for research purposes, and Apple has not announced any specific plans for integrating this into new or existing products. Still, it’s an interesting announcement from a company who no doubt has big plans for AI — and it’s another signal that in the future, one of the primary modes of engaging with software will be talking to it in natural language.
Thoughts on Gemini and How it Compares to GPT-4
Yes, you should be using it, at least side-by-side, for a while.
We’ve been kicking the tires on and reading about Google Gemini for a few days now, and are starting to form some initial opinions. Note that we are using (and paying for) Gemini Advanced, which uses Google’s “Ultra” model. Run-of-the mill Gemini also looks very powerful, but we prefer to use the leading models, and we also think this makes our comparisons to GPT-4 more apt.
Rather than detail our initial takes, we instead will quote Zvi Mowshowitz, who’s had access to Gemini Advanced for some time and sums things up as well or better than we could:
The biggest issue is that Gemini can be conservative with its refusals. It is graceful, but it will still often not give you what you wanted. There is a habit of telling you how to do something, when you wanted Gemini to go ahead and do it. Trying to get an estimation or probability of any kind can be extremely difficult, and that is a large chunk of what I often want. If the model is not sure, it will say it is not sure and good luck getting it to guess, even when it knows far more than you. This is the ‘doctor, is this a 1%, 10%, 50%, 90% or 99% chance?’ situation, where they say ‘it could be cancer’ and they won’t give you anything beyond that. I’ve learned to ask such questions elsewhere.
There are also various features in ChatGPT, like GPTs and custom instructions and playground settings, that are absent. Here I do not know what Google will decide to do.
I expect this to continue to be the balance. Gemini likely remains relatively locked down and harder to customize or push the envelope with, but very good at normal cases, at least until OpenAI releases GPT-5, then who knows.
There are various other features where there is room for improvement. Knowledge of the present I found impossible to predict, sometimes it knew things and it was great, other times it did not. The Gemini Extensions are great when they work and it would be great to get more of them, but are finicky and made several mistakes, and we only get these five for now. The image generation is limited to 512x512 (and is unaware that it has this restriction). There are situations in which your clear intent is ‘please do or figure out X for me’ and instead it tells you how to do or figure out X yourself. There are a bunch of query types that could use more hard-coding (or fine-tuning) to get them right, given how often I assume they will come up. And so on.
That’s our experience as well.
We also will note that Google seems to save all your chat data for several years by default. From TechCrunch:
Switching off Gemini Apps Activity in Google’s My Activity dashboard (it’s enabled by default) prevents future conversations with Gemini from being saved to a Google Account for review (meaning the three-year window won’t apply). Individual prompts and conversations with Gemini, meanwhile, can be deleted from the Gemini Apps Activity screen.
But Google says that even when Gemini Apps Activity is off, Gemini conversations will be saved to a Google Account for up to 72 hours to “maintain the safety and security of Gemini apps and improve Gemini apps.”
“Please don’t enter confidential information in your conversations or any data you wouldn’t want a reviewer to see or Google to use to improve our products, services, and machine learning technologies,” Google writes.
So manage your data and privacy accordingly. As for us, we’re not putting anything into Gemini that’s not anonymized, at least until they offer some enterprise-level security (as we have with our instances of GPT-4).
To give some sense of Gemini, we’ve conducted a set of side-by-side queries with Gemini Advanced and GPT-4, from the mundane to the more sophisticated. We post them here so you can judge them for yourself. In all cases we used the stock technology with no custom instructions.
Query: You're an audiophile an expert in turntables. How do I align my phono cartridge?
GEMINI:
Absolutely! Aligning your phono cartridge is essential for great sound and the longevity of your records. Here's a detailed breakdown of the process:
Tools Needed:
Protractor: Cartridge alignment protractors come in various forms (two-point, arc, etc.). You can buy one, find a printable one online, or some turntables even include them.
Small Screwdriver: To loosen and tighten the cartridge screws.
Tiny Ruler or Caliper: (Optional, but ideal) To precisely measure overhang adjustments.
Good Lighting and Steady Hands
Procedure:
Mount the Cartridge (if not already): Carefully attach your cartridge to the headshell on your tonearm, leaving the screws slightly loose to allow for adjustments.
Set Approximate Overhang: Overhang is the distance from the tonearm's pivot point to the stylus tip. A rough starting point is fine here – your protractor sometimes aids with an initial guideline.
Position the Protractor: Secure the protractor on your platter with the spindle through the middle. Double-check its levelness.
Align to the Outer Grid:
Move the tonearm so the stylus sits over the outer grid's marked point.
Gently adjust the cartridge within the headshell until its sides and front are parallel with the grid lines.
Align to the Inner Grid:
Pivot the tonearm to place the stylus over the inner grid point.
If the cartridge maintains parallel alignment with the grid lines, great! If not,minor adjustments to overhang are needed. Move the cartridge forward or backward in the headshell in tiny increments.
Repeat as Needed: Go back and forth between the outer and inner grid points until both show accurate alignment. This iterative process is expected.
Tighten and Verify: When both points on the grid align, carefully tighten the cartridge screws. Do a final alignment check on both grids in case slight cartridge shifts occurred while tightening.
Key Points:
Specific Cartridge Alignment Geometry: Protractors are designed for specific alignment methods (Baerwald, Lofgren A/B, Stevenson). Follow your protractor's instructions for the best results.
Take Your Time: This is a process requiring patience and fine adjustments. Improper alignment leads to distortion and uneven record wear.
Tonearm Setup: Before alignment, ensure the following:
Proper tracking force set according to your cartridge's recommendation.
Anti-skate adjusted to roughly match the tracking force.
Additional Tips:
Invest in a magnified headshell loupe. It helps enormously to verify precision at those grid points.
If your tonearm allows it, slightly offset the cartridge towards the spindle of the record for the inner grid (a few degrees). This technique minimizes distortion across most of the playing surface.
Seek out videos demonstrating the process if you are a visual learner.
Let me know if you have questions about your specific turntable or cartridge - happy to tailor the advice further!
GPT-4
Aligning your phono cartridge is crucial for achieving the best sound quality and ensuring your records play smoothly without unnecessary wear. Here's a step-by-step guide on how to align your phono cartridge:
1. Gather Necessary Tools
Alignment Protractor: A tool to help you accurately position the cartridge.
Screwdriver: For adjusting cartridge screws.
Magnifying Glass: To better see the alignment.
Tweezers: Optional, for handling small parts.
2. Loosen the Cartridge Screws
Using your screwdriver, slightly loosen the screws that hold the cartridge in place on the headshell. You want it secure enough to stay in place but loose enough to adjust.
3. Choose an Alignment Protractor
There are several types of alignment protractors, such as Stevenson, Baerwald, and Loefgren. The choice depends on your preference and turntable requirements. Baerwald is a common choice for its balance between inner and outer groove distortion.
4. Position the Protractor
Place the alignment protractor on the turntable platter. Ensure the turntable is completely level before proceeding. The protractor will have a null point or points where the cartridge should be perfectly aligned.
5. Align the Cartridge
Carefully adjust the cartridge's position so that the stylus tip sits perfectly on the specified point or points on the protractor. The cartridge body should be parallel to the protractor grid lines when viewed from above. Use a magnifying glass to ensure precise alignment.
6. Adjust the Overhang
The overhang is the distance the stylus tip extends beyond the center of the spindle. Adjust the cartridge in the headshell so that the stylus tip aligns with the protractor's guideline for overhang.
7. Tighten the Screws
Once the cartridge is properly aligned, carefully tighten the screws without shifting its position. Do not over-tighten, as this can damage the cartridge or headshell.
8. Double-Check Alignment
Recheck the alignment with the protractor to ensure that adjusting has not misaligned the cartridge. Make any necessary adjustments.
9. Set Tracking Force
After alignment, set the correct tracking force as specified by the cartridge manufacturer. This is crucial for proper playback and to avoid record damage.
10. Test Playback
Play a record to test the alignment. Listen for balanced sound across channels and absence of distortion. Minor adjustments might be needed for optimal performance.
Notes:
Patience is Key: Take your time with each step to ensure accuracy.
Consult the Manual: Some turntables or cartridges come with specific instructions for alignment. Always consult these first.
Aligning your cartridge can dramatically improve your turntable's performance and extend the life of your records. If you're unsure or uncomfortable with the process, consider seeking assistance from a professional.
We’d call that a tie. Now for something more pragmatic.
Query: Please proofread this text, listing any errors. Do not give me revised text. Just list the errors: [TEXT PASTED IN].
GEMINI:
Here's a list of errors found in the provided "Strategic Message Agenda" text. I've categorized them for clarity:
Grammar/Syntax
Inconsistent Verb Tense: The Communication Objectives section mixes present and future tense (e.g., "want to stakeholders" vs. "will strive").
Word Choice:
"Mike’asd" – likely a typo, needs correction.
"Linguistic Frame," "Organizational Metaphor," and "Narrative Form" might be overly jargonistic for some audiences.
Subject-Verb Agreement: "...as Mike has articulated it" should refer to 'vision' (singular) rather than 'company strategy and vision' (plural).
Punctuation
Missing Commas: Several places need commas for proper sentence structure or to separate items in a list.
Example: "...information and answering questions".
Formatting
Inconsistent Bullet Points: Some sections use circles (•), others use dashes (•).
Indentation: The "Key Things to Say Now" section has inconsistent indentation under "How we will achieve it".
Content/Clarity
Ambiguity: Some statements could be made clearer and more concise.
Example: "Our leader is transparent and sincerely listens" – Is this referring to the ORG leader or another figure?
Audience Consideration: Depending on the specific audience (internal employees, franchisees, etc.), some messaging might need revision for appropriate tone and understanding of company-specific terms.
Additional Notes
Overall, the Strategic Message Agenda is well-structured. However, minor errors can reduce its professionalism.
It's worth reviewing the document with the intended audience in mind to ensure all of the language resonates clearly.
Let me know if you'd like any specific examples of where these errors occur within the text!
GPT-4:
Here is a list of errors found in the text:
"providingas" should be "providing as" (typographical error).
"want to stakeholders and audiences to:" is awkwardly phrased and may be missing a verb between "want to" and "stakeholders".
"know or better understand;" should end with a colon instead of a semicolon for consistency with list introductions.
"We need localization but the individuality and competitiveness across markets doesn't serve the company as a whole" is a complex sentence that could benefit from clearer punctuation or restructuring for clarity.
"This is a symbol of the Mike’asd" seems to contain a typo or error ("Mike’asd").
"Confident in Our leader’s leadership" inconsistently capitalizes "Our".
"The singulared message" should probably be "The singular message" or "The single message".
"an essential part our strategy" is missing "of" ("an essential part of our strategy").
"makes it harder for us to succeed." This bullet point is correctly structured but lacks parallelism with the other bullet points which start with "This is/means".
The document's formatting and punctuation are generally consistent, but careful review for consistency in list punctuation and the use of semicolons vs. colons would be beneficial.
This list focuses on typographical errors, awkward phrasing, and potential areas of improvement for clarity and consistency.
A clear win for GPT-4 on proofreading. Now for some creative work:
Query: You're a seasoned PR professional. Please draft a press release announcing the opening of a new manufacturing facility for WidgetCo in the Philadelphia Naval Yard. It will employ 130 people, and will make step-down transformers use in graphic processing units. It will use clean energy. WidgetCo is owned by a non-US business, so we expect some to have a "not invented here" reaction. Include a quote from the CEO, Diane Jones: "We could not be more excited about our new relationship with the city of Philadelphia, and we're excited to bring both jobs and leading technology to this critical region." Include a quote by the CEO of the Naval Yard, Mike Shipsburg: "WidgetCo is just the type of business we've renovated the ship yard to host, and we know their interest signals a new era for the yard and the working people in this community." Presume the audience is a skeptical audience. Avoid pablum and PR cliches.
GEMINI:
FOR IMMEDIATE RELEASE
Contact: [Your Name] [Your Title] [Your Email]
WidgetCo Opens New Clean Energy Manufacturing Facility in Philadelphia Naval Yard, Creating 130 Jobs
PHILADELPHIA, PA – February 10, 2024 – WidgetCo, a leading global manufacturer of specialized components for advanced electronics, today announced the opening of its new state-of-the-art manufacturing facility in the Philadelphia Naval Yard. The facility will focus on the production of step-down transformers, critical components used in graphic processing units (GPUs). This expansion into the U.S. market will create 130 new jobs in the Philadelphia region.
The new facility is a significant investment for WidgetCo and reflects the company's commitment to sustainable manufacturing practices. The plant will be powered entirely by clean energy sources, demonstrating WidgetCo’s dedication to environmental responsibility.
"We are thrilled to be establishing a presence in Philadelphia, a city with a rich history of manufacturing and innovation," said Diane Jones, CEO of WidgetCo. "This new facility will allow us to better serve our growing customer base in the United States while creating high-quality jobs in the region. We could not be more excited about our new relationship with the city of Philadelphia, and we're excited to bring both jobs and leading technology to this critical region."
WidgetCo's commitment to the Philadelphia community extends beyond job creation. The company plans to collaborate with local universities and technical colleges to develop workforce training programs and ensure a pipeline of skilled workers for the region's manufacturing sector.
“We welcome WidgetCo to the Philadelphia Naval Yard with open arms,” said Mike Shipsburg, CEO of the Philadelphia Naval Yard. “WidgetCo is just the type of business we’ve renovated the shipyard to host, and we know their interest signals a new era for the yard and the working people in this community.”
About WidgetCo WidgetCo is a leading global manufacturer of specialized electronic components [add a bit about the company's established history if possible]. The company is committed to innovation, sustainability, and providing its customers with the highest quality products and services.
Let me know if you'd like any changes or adjustments to this draft! Here's why I think it works for a skeptical audience:
Focus on Facts: Emphasizes job creation, clean energy, and potential workforce development. This appeals to those concerned about economic growth and environmental responsibility.
Credible Quotes: Quotes from both CEOs lend legitimacy and highlight the mutual benefits of the project.
Addresses "Not Invented Here": Diane Jones' quote directly acknowledges the international aspect while framing it around economic benefits and local relationships.
No Fluff: Avoids overused PR phrases and focuses on tangible information.
GPT-4:
FOR IMMEDIATE RELEASE
Philadelphia, PA – WidgetCo, a leading manufacturer in the electronics sector, is proud to announce the opening of its new state-of-the-art manufacturing facility located in the historic Philadelphia Naval Yard. This strategic expansion is set to significantly bolster the local economy by creating 130 new jobs and introducing innovative manufacturing techniques for step-down transformers used in graphic processing units. WidgetCo, while rooted internationally, is committed to sustainable development and will power its operations entirely through clean energy sources.
The decision to set roots in Philadelphia represents a pivotal moment for WidgetCo and the city alike. CEO Diane Jones expressed her enthusiasm, stating, "We could not be more excited about our new relationship with the city of Philadelphia, and we're excited to bring both jobs and leading technology to this critical region. Our commitment goes beyond business growth; it's about fostering community development and environmental stewardship."
The Philadelphia Naval Yard, under the visionary leadership of CEO Mike Shipsburg, has undergone significant renovations to attract cutting-edge businesses like WidgetCo. Shipsburg shared, "WidgetCo is just the type of business we've renovated the shipyard to host, and we know their interest signals a new era for the yard and the working people in this community. Their presence here is a testament to the yard's potential as a hub for technological innovation and economic renewal."
The opening of WidgetCo's facility is not just a win for Philadelphia but a beacon of progress, showcasing how international collaboration can lead to shared prosperity. While some may harbor a "not invented here" sentiment towards foreign-owned companies, WidgetCo's investment in the Philadelphia Naval Yard exemplifies how cross-border partnerships can bring about mutual benefits, including job creation, technological advancement, and environmental sustainability.
WidgetCo is dedicated to becoming an integral part of the Philadelphia community, contributing to its economic vitality and technological landscape. The company looks forward to a long and fruitful relationship with the city and its people, proving that innovation, sustainability, and community engagement can go hand in hand.
For further information, please contact: [Your Contact Information]
This press release marks a significant milestone for WidgetCo and the Philadelphia Naval Yard, emphasizing the positive impacts of international investment in local economies and the shared vision of a sustainable future. WidgetCo's commitment to clean energy and job creation sets a precedent for future developments within the Naval Yard and beyond.
Note Gemini’s habit of explaining itself and trying to be helpful at the end of a response. Both outputs are reasonable drafts — certainly good enough to be a strong starting point for human hands. Gemini provided a more fulsome template, though, and note that it added additional quotes beyond the ones we provided in the prompt (and they work, actually). It also completely hallucinated a workforce training program, however. This is a big strike against it, in our view, because if the human using this were asleep at the wheel, they could easily miss this hallucination, which could create significant reputational risk for the organization. In terms of tone, both work, though we find the Gemini text a bit easier on the ears than that of GPT-4.
One more side-by-side:
Query: I am going to paste is a set of context to help you craft a narrative about a topic: communication principles that should guide communication about the topic; communication objectives of what we would like the audience to do, believe, know, and feel; framing guidance for the language to use when communicating about the topic; core messaging about the topic; and a narrative form that any narrative about the topic should honor. Based on this context, write a five paragraph narrative about the topic. The audience is employees in a fortune 100 organization. All are professional workers. They are open to new ideas but seasoned and have good and bad history with change. Here's the context for you: [THREE PAGES OF TEXT PASTED IN]
GEMINI:
Here's a five-paragraph narrative about ORG, crafted under the guidelines presented, and honoring the "Evolution" narrative form aimed at seasoned Fortune 100 professionals:Paragraph 1 In today's COMPANY, change is the only constant. It fuels our ambition, our relentless drive toward the One COMPANY Way. You've seen transformations before, some more successful than others. But ORG isn't another reshuffling or an empty rebranding exercise. It's about evolving how we work at the most fundamental level – streamlining how the entire COMPANY machine operates, ensuring all parts of the company function as a single, unstoppable unit.
Paragraph 2 Think of it this way: we're a highly capable team, built for a different game. Today, success needs something more. Sure, we excel individually, but it's time to become unbeatable together. This evolution will make life at COMPANY simpler, easier. Employees, suppliers, all of us—will find it easier to do the extraordinary things we're known for. Bureaucracy and outdated ways of working won't hold us back.
Paragraph 3 The team forming within ORG knows this journey holds challenges. The old guard sees echoes of past failures, some have understandable skepticism. We’ll meet that honestly. Our leadership’s committed to listening, to transparency, to finding better ways at every step. This won't be frictionless, but we'll earn your trust because this isn't a choice, it's a necessity for achieving our shared ambitions.
Paragraph 4 Imagine: A COMPANY where internal struggles vanish. One where every teammate seamlessly supports the broader mission. This evolution of ORG unlocks that unified power. Every market benefits, while simultaneously boosting global strength. That doesn't mean losing local expertise – quite the opposite – it enhances our responsiveness. Think of it as 'glocal': global strength built on a foundation of local knowledge.
Paragraph 5 There's a place for you within this new ORG, as partners shaping the very nature of how this company goes to market. Your voices, your talent, your experience – that's what builds a better future for us all. Evolution isn't comfortable, but it is how we win. Join us, and you'll be part of the team that creates an unbeatable, results-driven COMPANY. Together, we’ll succeed.
GPT-4:
In the journey toward excellence and innovation, COMPANY stands at a pivotal crossroads, embarking on a transformative path that will redefine our collective success. At the heart of this ambitious endeavor is ORG, a beacon of change and a testament to our unwavering commitment to our strategy, as envisioned by our leader, Mike. This initiative is not merely a shift in operations but a profound evolution of our organizational identity, aimed at enhancing collaboration, efficiency, and outcomes across every facet of our enterprise. We are poised to reimagine how we work, transcending past practices to meet our future aspirations head-on.
As we navigate this transformative journey, our communication principles serve as our compass, guiding us toward open, honest, and engaging interactions. By committing to transparency, simplicity, and genuine dialogue, we aim to foster a deeper understanding and appreciation for the pivotal role ORG plays in our collective success. This is a journey of significant change, one that requires us to reassess, reimagine, and reinvent how we operate. We acknowledge the challenges and resistance that accompany change, yet we approach this evolution with optimism, ready to embrace the opportunities it presents for growth and improvement.
Our objectives are clear: to forge productive partnerships, to inspire engagement, and to drive home the importance of ORG within the broader scope of COMPANY's strategy. This is not about maintaining the status quo but about breaking free from it. We recognize that the way we've worked in the past will not suffice for the future we envision. Through collaboration and a unified team effort, we will address and eliminate the friction that has hindered our progress, making it easier for all stakeholders to achieve their goals. This evolution represents a significant shift in how we perceive and engage with ORG, marking a departure from its current scope and reputation toward something far more impactful and valued.
The framing of our message around "results" and the metaphor of a "team" underscores our commitment to success and unity. We are not just colleagues; we are teammates, each playing a crucial role in the success of our collective mission. The emphasis on making it easier for all to succeed encapsulates our core message, focusing on the tangible benefits and improvements that ORG will bring to every level of COMPANY. This is about simplifying processes, enhancing experiences, and ultimately, achieving superior outcomes for our customers, suppliers, and ourselves.
In embracing this evolution, we stand on the threshold of a remarkable opportunity—an opportunity to be part of something truly groundbreaking. As we put our plan into action, we do so with a commitment to transparency, excellence, and an unwavering belief in our ability to succeed. This journey may be challenging, and not everyone will greet it with enthusiasm, but we are driven by the promise of a better tomorrow. We are good today, but through the evolution of ORG, we will be even better tomorrow, ensuring COMPANY's place as a leader in innovation and success. Together, as a unified team, we will navigate this journey, adapt, and emerge stronger, more capable, and more competitive than ever before.
We give the nod to Gemini here — which surprises us. To date, we’ve found GPT-4 (and sometimes, Claude 2), the only large language models capable of taking a body of context and crafting cogent narratives from it. While GPT-4 can do this very well, note in this example that it confounded elements of the context that should inform the narrative with information to convey in the narrative. And while the Gemini messaging is a bit over the top, it’s punchier, better adheres to a narrative form, and says more with less.
So what does it all mean? That’s hard to say, as there are dozens of use cases we’ve not put to comparison yet (for example: act as a critic, revise this text, theme these comments, summarize these notes, play devil’s advocate, brainstorm ideas, etc. etc.). In the least, though, it means you should probably be using both side-by-side, at least if you’re serious about understanding where the models are going and getting multiple takes on perspectives and output. Eventually, nobody is probably going to pay $20 or $30 a month to multiple LLMs that all presume to do the same thing. The open question is how similar these models are when they scale, versus how differentiated they can be. So far, that’s unknown, and the only way to figure out the similarities and differences is to use the tools. We will, as always, share more about our own experience in doing so as we learn more.
We’ll leave you with something cool, and perhaps a bit chilling: rooms full of working and autonomous human-like robots …
AI Disclosure: We used generative AI in creating imagery for this post. We also used it selectively as a creator and summarizer of content and as an editor and proofreader.