> That same government hasn't exactly managed any of its semi-public companies particularly well: the national telco is for shit, postal service is nearly bankrupt, railways are mismanaged and underfunded, etc.
In fairness, it's not the same gov that nuked the public service than the one in power now. But on the flip side, the selloff of public services to private sector was a success and achieved the stated goals: Destroy it from the inside and use that as an excuse for more liberalization.
Belgium's reactors are really old, and have lots of issues. They have been dragging their feet for decades on the subject and instead of building new reactors 10-20 years ago, they are now un-decomissioning older reactors..
> Belgium's reactors are really old, and have lots of issues.
I want to point out that Belgium has the (global) gold standard of nuclear regulation. They have annual reviews, 5 year major reassessments, and 10 year Periodic Safety Review (PSR). The purpose of the PSR is to build a plan to keep all nuclear plants up-to-date with state of the art safety mechanisms. Each PSR has mandatory upgrades. If operators fail or refuse these upgrades, they are forced to shutdown. There is no one other country who does nuclear safety quite like Belgium.
The incentive previously was having more secure software making a name for yourself. The incentive now is finding the most noisy vulnerability so you can push FUD to sell your AI software.
It used to be done for fame and visibility. Give a marketable name and a website, your exploit will be talked about and your name will shine in the industry.
Now it's done by an LLM to sell more LLMs services. Disclosure is botched to have the most sensational title so more click more upsell.
I'm being very cynical here but who says that their tool or LLM discovered this. How do we know they didn't hire some expert security researchers to find it or bought it off the black market as a promotion stunt.
With that being said, I wouldn't mind if they made more sales on whatever they're advertising IF they followed the disclosure process well. A bad disclose immediately tells me I can't trust them because their moment in the light was more important that the safety of millions of boxes.
They did not, in fact, botch anything. They notified the responsible party and followed a practice that is pretty much the accepted norm (and for good reason).
How recursive should their notifications be? Just the tip three distros? The top dozen? Every embedded Linux router company? How about every hosting provider?
They did what they're supposed to without being paid for it. The only other good source of funding for security research besides marketing budgets for security companies will NOT result in a disclosure timeline you'd be happier with. ;-)
It didn't seem that deep to me. They just saw an issue with Goblins, dissected the word from the model, then it appeared again in the next version without them knowing exactly how or why.
Goes to show it's all vibes when making these models. The fix is literally a prompt that says not to talk about goblins...
> We retired the “Nerdy” personality in March after launching GPT‑5.4. In training, we removed the goblin-affine reward signal and filtered training data containing creature-words, making goblins less likely to over-appear or show up in inappropriate contexts. Unfortunately, GPT‑5.5 started training before we found the root cause of the goblins.
The prompt is just a short term hotfix/hack because they couldn’t get the proper fix in in time.
If you need to put baby guardrails on your model because the training is effed up, maybe you should rethink how you make these models and how much control you really have on it.
Yes that was the primary issue I had when testing Zed in the past. Keyboard layout not working properly, shortcuts being unusable or un-remappable. Sad to see it's still the case for 1.0
In fairness, it's not the same gov that nuked the public service than the one in power now. But on the flip side, the selloff of public services to private sector was a success and achieved the stated goals: Destroy it from the inside and use that as an excuse for more liberalization.
reply