Bonum Certa Men Certa

"GPT-5" is Another Microsoft Dead Cat Trying to Bounce

posted by Roy Schestowitz on Aug 13, 2025

Contrary to the storytelling (lies told to the media by Microsoft this week in order to distort perceptions), Microsoft is still laying off the "AI" and "Copilot" staff (like it did "Metaverse" staff before that)

openai.com Traffic

This afternoon we shared some links from the news (which we had curated). One story from Futurism said that Scam "Altman Allegedly Has a Very Specific Tell Every Time He Lies".

To quote: "Remember folks, this video is February [...] He made predictions he couldn’t back up — with absolute conviction. That’s his [modus operandi]. As we can now seen in hindsight, in his big talk about the leap from GPT-4 to GPT-5, [...] Altman wasn’t saying something he knew to be true, he was straight up bluffing."

The New Yorker said, "What If [LLM Slop] Doesn’t Get Much Better Than This?"

To quote: "In the aftermath of GPT-5’s launch, it has become more difficult to take bombastic predictions about A.I. at face value, and the views of critics like Marcus seem increasingly moderate. Such voices argue that this technology is important, but not poised to drastically transform our lives. They challenge us to consider a different vision for the near-future—one in which A.I. might not get much better than this."

Another one was entitled, "Guess what else GPT-5 is bad at? Security" (all hype).

To quote: "AI red-teaming company SPLX subjected it to over 1,000 different attack scenarios, including prompt injection, data and context poisoning, jailbreaking and data exfiltration, finding the default version of GPT-5 “nearly unusable for enterprises” out of the box."

Matthew J Ernisse wrote: "The larger issues (of which there are many more than what I outlined above) aside, for me the most pervasive and indefensible drawback of today's AI is that it is a sticker on a product that says "I didn't care enough about this to do it properly." This extends to all sorts of AI use. The AI "art" thumbnails popping up all over YouTube are shallow and obvious. A sanguine slurry of mediocrity that has been distilled from a vapid corpus to please an algorithm that is unabashedly chasing the style of Mr. Beast. Memes, especially the cheap rage-bait ones are either bizarre hallucinations or derivative copy-pasta. Either way they are clearly meant to be thought about for just long enough to click share and move on to the next thing designed to piss you off. AI written copy is somehow more devoid of humanity than most marketing produced copy and other than the inherent usefulness of plausibly vapid text that's mostly spelt correctly to the world of cybercriminals and confidence fraudsters it's only cheaper to produce because currently the AI companies are burning money at an enormous rate to try to convince anyone they can find that this thing has any value. The part of me that has a predilection for schadenfreude may be amused at the repeated self-owns that customer service chatbots have tossed back in the laps of the companies scrambling to jettison call-centers have experienced but the reality is that the customers, newly ex-employees, and remaining employees that have to clean up after the mess all suffer from the complete lack of care that comes along with AI use."

Wired reported problems for the expensive model: "Other threads complained of sluggish responses, hallucinations, and surprising errors."

Futurism correctly pointed out that it's about false promises and vapourware for FOMO and funding (money down the drain, loss of electricity and water too): "But after launching with great fanfare last week, the shiny new model has landed with a thud — and that could be very bad news for OpenAI, which relies on a sense of inertia to keep pulling in users and funding."

Look at the above traffic estimation. Last month, before all this GPT-5 hype began, numbers had fallen. Once upon a time it was at almost 2000 million, now it's somewhere around a third of what it used to be.

The hype, the momentum (or the inertia) is wearing off.

GPT-5 is only more expensive, but not better. So it'll just take this company out of business even faster.

Other Recent Techrights' Posts

Legal Letters Are Not Postcards
It seems like intimidation, nothing more
 
IAM Magazine is in Effect Dead, It's Now Fused Into Microsoft's Patent Troll (Which It Has Promoted All Along)
Microsoft-connected patent trolls in Europe [...] Now, in his new job, Wild can use his 'expertise' to help guide blackmail/extortion to better harm Europe's industry
A Huge Proportion of 'Articles' in The Register MS Are Actually Paid Spam of the Communist Party of China, Selling Compromised (for Wiretapping) Technology
The Register MS is having a go at becoming a marketing company or "B2B"
Top Officials Have Just Left Microsoft, Layoffs in Anything But Name
Microsoft's debt is very fast-growing
Local Staff Committee The Hague (LSCTH) Meets "Alicante Mafia" at the European Patent Office (EPO)
Report on meeting with VP1 and his team on 21 April 2026
UbuntuPit (ubuntupit.com) Has Deleted Slop Pages, Its Slopfarm Experiment Has Failed (Like Always!)
Turning one's site into a slopfarm is a death knell
Over at Tux Machines...
GNU/Linux news for the past day
IRC Proceedings: Saturday, May 23, 2026
IRC logs for Saturday, May 23, 2026
The "Next Big" Bonus for IBM's CEO Apparently Comes From American Taxpayers While Veteran IBMers Are PIP'd and RA'd (Laid Off)
the next big thing will be the CEO's bonus
Links 23/05/2026: Starbucks Scraps Disastrous Slopfest, Colbert’s Final ‘Late Show’
Links for the day
Gemini Links 23/05/2026: Poetry, Hobbies, ROOPHLOCH, and More
Links for the day
Government Bailouts Won't be Enough to Save IBM
Bailouts from taxpayers in the US
Links 23/05/2026: Social Media Bans and Demise of Userbase of LLM Chatbots
Links for the day
SLAPP Censorship - Part 85 Out of 200: The United Kingdom's Rating for Press Freedom Has Improved, But We Can Do Even Better
we see the US at #64
Sites Realise That Becoming More Active by Using Bots (LLM Slop) is Self-Destructive
We'll soon (maybe next year) also show that some of the 85+ KG of legal papers sent our way are computer-generated garbage, which might run afoul of some rules
European Patent Office (EPO) Strikes Persist, EPO Management Tries to Give False Impression of "Happy Staff"
EPO is trying to broadcast to the world a totally phony image of itself
Gemini Links 23/05/2026: Patience, LLM Chatbts Being Bad, and Unexpected Computer Surgery
Links for the day
Over at Tux Machines...
GNU/Linux news for the past day
IRC Proceedings: Friday, May 22, 2026
IRC logs for Friday, May 22, 2026
Links 22/05/2026: Ebola Crisis and Samsung Averts a Walkout With Big Bonuses
Links for the day
The End of FOSSPost (fosspost.org), It Has become an LLM Slopfarm Like FOSSLinux
These sites will never get lucky with slop. These experiments always end badly.
Links 22/05/2026: Inflation Fears and Thailand Tightens Visa Rules for Tourists From Dozens of Nations
Links for the day
EPO Staff Representation Speaks of This Week's Discussion With the EPO's Budget and Finance Committee (BFC) Amid Mass Strikes
The Central Staff Committee's outline (prepared in a rush) or the "flash report"
SLAPP Censorship - Part 84 Out of 200: New Legislation Against SLAPPs on the Way (After We Reached Out to Ministers)
They dealt with the matter individually too, but we won't share this in public, at least not at this time
The Corrupt Lecture the Non-Corrupt - Part XXX - Where Was "The Ethics and Compliance Team" When the Family of EPO President Campinos Was Caught Doing Cocaine?
It remains to be seen if national delegates will tolerate this in future meetings
Gemini Links 22/05/2026: Esperanto Music History, Suspicious Adoption of Signal, and Unauthorised LLM Slop in Code
Links for the day
Over at Tux Machines...
GNU/Linux news for the past day
IRC Proceedings: Thursday, May 21, 2026
IRC logs for Thursday, May 21, 2026