I’m curious about why xAI’s Grok, with ~30 million monthly active users (MAUs), remains absent from DEVONthink 4.0’s AI lineup, which includes ChatGPT, Claude, and Perplexity. Grok’s conversational strengths and real-time X data access could add value to our projects, yet DEVONtechnologies’ responses on this topic have felt vague, leaving me wondering about the real reasons. I’d like to explore potential factors—politics, media narratives, or technical hurdles—and hear your perspectives on integrating Grok.
Are Political Perceptions at Play? Grok’s “truth-seeking” mission has stirred debate, with some media citing bias concerns (e.g., unverified claims linking it to controversial narratives). ChatGPT (~600M MAUs) and Claude (~30M MAUs) have faced similar scrutiny but are integrated. Could political sensitivities around Grok be a factor in DEVONthink’s hesitation? I’d value clarity on whether these perceptions influence the decision.
Do Media Narratives Around Musk or DOGE Matter? Grok’s ties to Elon Musk and X invite media scrutiny, with narratives sometimes framing Musk or DOGE (the cryptocurrency) skeptically. X posts have suggested Grok’s algorithm faces bias accusations, though evidence is limited. Might these media portrayals shape DEVONthink’s choice to prioritize other models? I’m interested in whether the community sees this as a barrier.
What About Grok’s 30M MAUs? Grok’s ~25–30 million MAUs rival Claude and Perplexity, and its API (launched April 2025) supports tasks like text generation, suitable for DEVONthink’s summarization or querying needs. While its cloud-based nature might raise privacy questions, DeepSeek’s inclusion shows flexibility. Given Grok’s scale, why isn’t it in the mix? Past responses from DEVONtechnologies haven’t shed much light, so I’m hoping for a more detailed explanation.
DEVONthink is a cornerstone for our workflows, and Grok could make it even better. Do you think Grok’s X data access or conversational style would benefit your work? DEVONtechnologies, could you provide a clearer picture of why Grok hasn’t been integrated? Is it politics, media, technical challenges, or something else? The brief answers so far haven’t fully addressed this, and I’d appreciate deeper insight.
Let’s share ideas and keep pushing to enhance DEVONthink’s AI options.
We get requests for all manner of integrations, many of which you never see or hear about as the forums aren’t our sole client interface. We can’t just slap in an endpoint and say, “Well, there you go!” It takes development and testing to see how things are working and to ensure they’re safe, private, and useful within the context of our applications. We have added very specific services based on time to market and user inquiry, with ChatGPT logically being the first in line. The others included predate Grok3. And honestly, we have not had many requests for Grok3 either.
Remember we are trying to move toward the gold master release of version 4. This is bug-fixing (working with very few actual reports from people) and polishing time.
Could it be included in the future? Possibly, but it depends on timing, interest, and available resources.
Thank you for addressing my inquiry about Grok’s integration into DEVONthink 4.0. While I value the challenges of your development schedule, I found the response’s tone and lack of depth disappointing.
The phrasing, such as “We can’t just slap in an endpoint,” implies limited technical understanding on my part—a presumption that is misplaced, given your lack of knowledge about my expertise. A neutral tone would have been more suitable.
Moreover, the reply bypasses my questions on political sensitivities surrounding Grok, media narratives linked to Elon Musk, its equivalence to Claude and Perplexity, and the potential of its X data access. Addressing these would have enriched community dialogue.
As a dedicated DEVONthink user, I sought a response matching my inquiry’s rigor. Enhanced transparency would foster greater trust and collaboration.
I do fully comprehend that the inquiries pertaining to any conceivable political sensitivities involving Grok and Mr. Musk hold a matter of considerable import to your good self. Nevertheless, whilst each and every one of us here at DEVONtechnologies undoubtedly harbours our own individual and diverse political convictions, the corporation as a collective entity—being but an assembly of such individuals—possesses, I dare say, but scant political inclinations, if any at all. Consequently, it is our firm resolve to refrain from offering commentary on political affairs or to espouse any particular faction, should such matters arise.
Moreover, I am duly mindful of the regard in which you hold Grok, having elected it as your preferred AI provider. As Jim has so aptly observed, there exist a multitude of such models, with new ones burgeoning like mushrooms. Our resources, though earnest, are finite, and we allocate them in a manner we judge most conducive to the fulfilment of our mission. Whilst it is conceivable that this posture may evolve in times to come, we are wont to abstain from discourse on prospective designs, for such intentions are ever subject to alteration by forces both within and beyond our control.
With regard to the delicate matter of trust, I find myself most humbly disinclined to presume upon such a weighty question, and thus shall defer to the better judgment of others more suitably qualified to pronounce thereon.
I find BLUEFROG’s tone completely neutral, and I don’t see how it implies anything about your technical understanding.
According to xAI’s release notes, the Grok 3 API launched on April 3rd—the exact same day as DEVONthink 4.0beta1. It’s only been 2 months. That timeframe is way too short to assume anything in my opinion.
I personally have no desire to use any of these LLMs, so I don’t care either way. And I’m not informed enough to have an opinion worth sharing regarding the merits of any particular model.
I’m not sure what threshold must be met before you find something transparent enough, but I don’t think it’s reasonable to expect DEVONtech to make all their internal deliberation public.
I rather read @eboehnisch answer as a statement to choose not get dragged into the modern (or en vogue) political play of avowals, pressed or otherwise.
I think that there is a right ‘not to state’, esp. for a company, when everybody knows these answers can easily be fed into a pinball discourse machine…
… JMO; and, granted, maybe I read too much Foucault…
It’s been made clear DT isn’t going to get into political/policy responses. And what for? So people can complain when the answer invariably doesn’t comport with their views? It’s a lose/lose proposition. DT is an app we use to get our work done. Politics are irrelevant. Let it go.
Oh, I see I used the wrong reply button. My reply was @Deeptide, as was my request for answers to the questions posed by @Deeptide, which I think was a reasonable request.
I would say that it’s unwise for any corporate brand to become embroiled in politics, because choosing sides will inevitably make the other side unhappy.
That’s true of DevonTechnologies, but it’s even more true of Grok and xAI. Musk is a polarizing figure because of choices that he has made. xAI has political baggage – fairly or unfairly – because of choices made by its owner.
as was my request for answers to the questions posed by @Deeptide, which I think was a reasonable request.
No, not reasonable. Questions 1 and 2 were politically loaded. Not only that, but already in the title, @Deeptide is insinuating motives. Yet, they got quickly offended by the rather neutral “slapping an endpoint in” remark. By giving other, sensible reasons, Q1 and Q2 are implicitly answered. The post could have started out with a friendly, neutral, “Hey, any chance that xAI Grok will be added in the future”? People could have chimed in, voicing support. Instead we got this. As they say in German, “der Ton macht die Musik”. My answer: Grok will be added when the DT devs see fit.
Which is why I asked @Deeptide to answer them. Rather than hiding behind the language they used, I wanted them to come out and say exactly what they were getting at. “Narrative”, “portrayal”, words which I’m pretty sure are being used, here, as a euphemism for “lies”. So, yes obviously politically loaded. To be absolutely clear, my post above was directed to @Deeptide and not to DT. FFS.
@Deeptide was asking DT specifically about DT’s motivations. As such, OP cannot give an answer. Having said that, it is obvious that OP suspects the answers are “yes” to both. Unlike you, I will end this post without an expletive.