Chronic Lying as a Prerequisite and a Lifestyle of Proprietary Software Pushers (Who Now Promote to Us Lying Machines, LLMs, as "Intelligence")
A bunch of posers and pretenders like S[c]am Altman, calling proprietary "open", lies "intelligence", and an epic business failure (billions of dollars in losses) "revolutionary"
LIARS thrive in workplaces where lying is routine. Honest people thrive in places that are transparent and ethical. Since 2019 when Bill Gates sent some bribe under an NDA my last employer has been a place where lying was routine and Free software was being gradually dismantled. Managers neither used nor understood Free software. They just didn't care. Contracts were being illegally signed, staff was lied to, customers were lied to, and so on. It was terrible. I already wanted to leave.
This morning we entitled Daily Links "Horizon Proprietary Software (Linked to Microsoft Suppliers) Killed Innocent Post Office Workers" because it's all over the news here (since last week). We already see how code being secret leads to deaths, with notable examples like 'Dieselgate' and Boeing.
The thing is, based on my personal experience, people who try to sell secret code or compiled (binary-only) software are dodgy, manipulative liars. They know that what they sell is deeply defective and they hide all the problems by concealing the source code (denying access, even by independent auditors).
When you worship and mimic a "master liar" like a bobbing Bill Gates, lying about his crimes while interrogated in a room in front of a camera, perhaps lying to judges seems rather "normal". There was lots of that in "Horizon". Had only these people had a sense of guilt or consciousness, a lot of this disaster would be averted (and a lot earlier, too).
All Roads Lead to Rome Microsoft?
The hint as to the culprit remains buried behind a layer or a network of contractors, which include Microsoft's golden lovers like Accenture and Fujitsu. This is possible to trace back and demonstrate based on news reports.
The linking to Microsoft is not as important as the proprietary nature, an associate has explained. And it is the proprietary nature which in and of itself caused the problems. Then on top of that you have the lying and cover-ups.
At that point one can speculate and cast aspersions about it being linked to Microsoft while admitting that there is not yet evidence of a link beyond the Microsoft style failure, Microsoft style lying, and Microsoft style victim-blaming. However, it is probably more than sufficient to just rightfully criticise the nature of proprietary software.
How many people did this scandal kill so far? It ended not just the professional lives of many innocent people. It literally killed many.
A Cuppa Nothing
Speaking of dishonesty from Microsoft, last night we wrote about Microsoft's failure to catch on and to actually make a dime from LLMs. It's an extraordinary smokescreen while lots of layoffs are happening at Microsoft. They need a distraction and they need to bamboozle shareholders.
When it comes to Microsoft, the details remain obscured. It seems safe to assume that "content" "partnerships" Microsoft and 'Open' Hey Hi' (OpenAI) boast that they sign with publishers are actually pre-court settlements, with a price agreed upon upfront (Microsoft did the same when it blackmailed Linux OEMs with software patents).
They just use flowery language to hide the fact that on top of losing money on staff/electricity they also lose money in court. LLMs have no viable business model because they're so pointless (in practice) that nobody would pay for them.
An associate agrees on the point that 'plausible sentence generators' (LLMs) have no viable business model. They have no real use and thus no motive for people to pay for them. Again, this whole adventure is only possible due to 1) Microsoft marketeers 2) Azure being over-provisioned and underused and thus available for games like LLMs.
Ironically, a year later, a lot of the media talks of LLMs not from the framing or perspective Microsoft wants. The media, now being paid less by Microsoft, points out that medical diagnosis from ChatGPT is wrong in about 9 out of 10 cases, based on a study. This past week a lot of attention was paid to how ChatGPT fabricates court cases or precedents, albeit excusing those away as "hallucinations" rather than severe, inherent errors, highlighting the lack of reliability of LLMs. They're infeasible for most applications which they were advertised for, much like "Horizon". As a result, innocent people suffer while liars and plagiarists thrive.
As already noted in Daily Links, Microsoft employees continue to swarm the Board of "OpenAI" ('Open' Hey Hi'), in effect controlling the company without even buying it, as there's an M&A budget issue, necessitating an Elop/Nokia-like manoeuvre. Microsoft never plays based on the rules, not even based on the laws. █