In this AG Connect article: 'The Age of SaaS Is Over', the CFO of OpenAI suggests that SaaS is EOL! Because companies that embrace AI will develop their own software with it...
The first sting I see looming: How PROPRIETARY is that "proprietary software"? And do the "programmers" of that software have the knowledge to see if it's sound and what's happening (in the background) with YOUR data? Are the terms of use clear about this, and aren't they being changed secretly?
In addition, it seems to me UNTHINKABLE that a system, (Open)AI, that owes its right to exist and progress to an insatiable, ever-growing hunger for data and requires truly astronomical investments and energy to run, will turn out to be a more or less "free beer" or solution with a better ROI than SaaS services!
Considering the above, I DO understand that (Open)AI would also like to dig into the data from YOUR SaaS solution! Where it doesn't have access yet...
That is why we reiterate here once again the advantages of the open-source FAMP stack, WITHOUT "fast fashion" "proprietary" software, under which QUARAS offers all online SaaS services:
• No worries your data might (by silent terms and conditions) be subject to AI training! (ChatGTP, Copilot, Gemini, Grok or DeepSeek)
• No hidden costs for or coercion to purchase services you do NOT want!
• Anything and everything can be tailored specific to your needs!
• Serious alternative to mono culture one-size-fits-all online app landscape!
• and... BUILT ON THE RIGHT TOOLS FOR STABILITY, PERFORMANCE AND SAFETY!
We will of course always keep a close eye on all developments here, so hopefully until next time!
Consciously chosen to write THE word completely together, because according to this AG connect article: AI in healthcare stumbles over simple spelling errors. Now, I'm the first to openly admit that it's often nigh-impossible to decipher (handwritten) doctor's prescriptions and advice, and that even the fastest quantum computer ever developed will likely take longer to do so than the existence of the universe itself... But still, hopefully, you also see the fundamental problem this exposes, and as you know by now, I like to call a spade a spade, so here we go...
Namely, that AI (still) does NOT possess the "intelligence" to truly understand the subject matter! That is, to turn data and patterns into knowledge and insight! And without a significant structural adjustment, I NEVER see that happening. More data and patterns do NOT lead to broader understanding or higher-level insight. So the models will truly have to become "self-learning", and will NVIDIA then be able to develop the corresponding chips? Until then, AI is nothing more than a speed-reading aid that provides you with the (book) excerpt. But hey, if you read quickly, you occasionnally will miss some things.
That is why we reiterate here once again the advantages of the open-source FAMP stack, WITHOUT the costly, energy-consuming speed-reading aid, under which QUARAS offers all online services:
• No worries your data might (by silent terms and conditions) be subject to AI training! (ChatGTP, Copilot, Gemini, Grok or DeepSeek)
• No hidden costs for or coercion to purchase services you do NOT want!
• Anything and everything can be tailored specific to your needs!
• Serious alternative to mono culture one-size-fits-all online app landscape!
• and... BUILT ON THE RIGHT TOOLS FOR STABILITY, PERFORMANCE AND SAFETY!
We will of course always keep a close eye on all developments here, so hopefully until next time!
I read this AG connect article: KPMG writes a 100-page prompt to have an AI agent provide tax advice with increasing amazement! While the title was already hilarious, it got progressively worse as I went along...
A "substantial team" has been working on it for months, the taxbot cannot operate completely independently, and the output is so complex that it can only be processed by a certified tax consultant!!! Where/what is the ROI of Agentic AI here? This sounds suspiciously like the promise of "low" code applications that require constant support from an army of programmers... hopefully, you understand the paradox.
(Agentic) AI undoubtedly has its uses, but in my opinion, ONLY in marketing jargon that has to be hip, sexy, shiny, flashy, and fast, but doesn't actually have to be accurate at all! That's why at QUARAS we still see a bright future for traditional applications based on structured data (SQL tables) that always quickly and reliably deliver the right KPIs and answers to make informed decisions. So, always 100% reproducible results, just like we used to do science...
That is why we reiterate here once again the advantages of the open-source FAMP stack under which QUARAS offers all online services:
• No worries your data might (by silent terms and conditions) be subject to AI training! (ChatGTP, Copilot, Gemini, Grok or DeepSeek)
• No hidden costs for or coercion to purchase services you do NOT want!
• Anything and everything can be tailored specific to your needs!
• Serious alternative to mono culture one-size-fits-all online app landscape!
• and... BUILT ON THE RIGHT TOOLS FOR STABILITY, PERFORMANCE AND SAFETY!
We will of course always keep a close eye on all developments here, so hopefully until next time!
In this AG connect article by Jasper Bakker: Moving away from Java saves Apple 40% in performance, it becomes clear once again that "the right tool for the job" can have a huge impact. And then it is NOT just about performance, but also maintainability, flexibility, complexity and as a derivative therefore safety!
Standardization on one platform or programming language has its advantages, after all, you only have to learn one environment. But if that environment tries to be everything for everyone, it can bring so much overhead that it overshoots its goal. I'm talking about Java of course.
The reverse can also become problematic if the tool is too limited, for example Javascript for both client and server functionality. A single process environment in which you do a multi process(or) simulation via complicated constructions (promises). And because of the complexity you seek refuge in frameworks to keep the development manageable! With all the risks that you bring in (unseen) with that...
That is why we reiterate here once again the advantages of the open-source FAMP stack under which QUARAS offers all online services:
• No worries your data might (by silent terms and conditions) be subject to AI training! (ChatGTP, Copilot, Gemini, Grok or DeepSeek)
• No hidden costs for or coercion to purchase services you do NOT want!
• Anything and everything can be tailored specific to your needs!
• Serious alternative to mono culture one-size-fits-all online app landscape!
• and... BUILT ON THE RIGHT TOOLS FOR STABILITY, PERFORMANCE AND SAFETY!
We will of course always keep a close eye on all developments here, so hopefully until next time!
In this AG connect article by Chris Verhoef, reference is made to the days of yore when a range of different providers with their own (often unique) solutions competed in automation projects, which were all later revamped to tendering processes that got bogged down in stifling regulation, I add on personal title!
And there you have the birth of the uniform Big Tech IT landscape, because for one reason or another they ALL DO get through those tendering processes...
Was everything better in the past? Nostalgia distorts, so probably not! Is everything better now? Because it has evolved! I doubt it, self-interest always plays a major role in evolution and the "bigger" that self-interest, the less room for healthy competition! And the more unequal the relationship between the (Big Tech) supplier and (small) customer!
"Towards a world in which your data is truly yours and yours alone." is a final conclusion that we at QUARAS wholeheartedly endorse and hereby invite you to also take the associated step towards a "small Tech" provider.
That is why we reiterate here once again the advantages of the open-source FAMP stack under which QUARAS offers all online services:
• No worries your data might (by silent terms and conditions) be subject to AI training! (ChatGTP, Copilot, Gemini, Grok or DeepSeek)
• No hidden costs for or coercion to purchase services you do NOT want!
• Anything and everything can be tailored specific to your needs!
• Serious alternative to mono culture one-size-fits-all online app landscape!
• and... READY FOR QUANTUM COMPUTING!
We will of course always keep a close eye on all developments here, so hopefully until next time!
In this ISACA article courtesy to Rob Clyde, the claim is that there are no more excuses for procrastinating on quantum preparation!
Especially considering NIST finalized a set of post-quantum cryptographic (PQC) standards already available in Google Chrome based internet browsers as explained in this Forbes article.
But how difficult and cumbersome is it to implement this? Well... quicker than writing this short piece! Just a couple of steps needed as explained in this article by Bernard Spil:
• Install two (small) ports from the FreeBSD portstree
• Add three lines to openssl.cnf and the one line below to httpd.conf:
• SSLOpenSSLConfCmd Curves X25519MLKEM768:X25519:prime256v1:secp384r1
• Restart Apache et presto
To be fair, we have several vhosts for which we had to add the same line above...
Second side note, if you are asked to upgrade OpenSSL to V3.2+, do NOT choose V3.4, as it will cause problems with Apache!
That is why we reiterate here once again the advantages of the open-source FAMP stack under which QUARAS offers all online services:
• No worries your data might (by silent terms and conditions) be subject to AI training! (ChatGTP, Copilot, Gemini, Grok or DeepSeek)
• No hidden costs for or coercion to purchase services you do NOT want!
• Anything and everything can be tailored specific to your needs!
• Serious alternative to mono culture one-size-fits-all online app landscape!
• and... READY FOR QUANTUM COMPUTING!
We will of course always keep a close eye on all developments here, so hopefully until next time!
Officially, according to the article on AGConnect, the price increases are necessary for maintenance and development of updates...
Since the latter seems very unlikely, the basis undoubtedly lies elsewhere. And we can think of a plausible reason for it: A fine for OpenAI / Copilot inaccessibility of your data for "training purposes"!
If you want to have (parts of) your data unlocked via AI, it is nice that you retain control yourself and do NOT pay for something that you do NOT purchase! Not even indirectly.
That is why we reiterate here, as in our previous article, the advantages of the open-source FAMP stack under which QUARAS offers all online services, that can also run entirely on-premise:
• No worries your data might (by silent terms and conditions) be subject to AI training! (ChatGTP, Copilot, Gemini, Grok or DeepSeek)
• No hidden costs for or coercion to purchase services you do NOT want!
• Anything and everything can be tailored specific to your needs!
• Serious alternative to mono culture one-size-fits-all online app landscape!
We will of course always keep a close eye on all developments here, so hopefully until next time!
Here at QUARAS we strongly believe so and are providing our wide range of proven online solutions since 2004 on this open source software stack!
BTW: FAMP = FreeBSD + Apache + MariaDB + PHP
In upcoming articles we will highlight in more depth the many benefits and freedoms QUARAS can provide by NOT being locked in to the big players including for data storage:
• Serious alternative to mono culture one-size-fits-all online app landscape!
• No worries your data might (by silent terms and conditions) be subject to AI training! (ChatGTP, Copilot, Gemini, Grok or DeepSeek)
• Speaking of data security: HTTPS as single protocol for ALL client access!
• Anything and everything can be tailored specific to your needs!
Hope you will return for the in-depths next time, but meanwhile feel free to look around and contact us for any queries you might have!