WhatsApp started flashing an expanding "Ask Meta AI" tooltip every time you open the app recently.

I guess this is the moment when they realise that the only reason why some people use AI because it's free or forced upon us. No marketing campaign will ever make people pay this sort of money pay for a silly chatbot, especially the Meta AI crap.

2 days a week take your big fat arse out of your executive chair and sit with your ground floor emoloyees. Try to reconnect. Understand what they do. How they do it? Why do they do some things in a particular way, and why they don't do certain things at all. Prick your ears and listen to what happens in the office - praise, but also complaints - large and small. Maybe even enroll yourself in an entry level course on whatever training platform you pay for. Reconnect. Learn. Understand. Then, and only then make decisions. Only when you understand what you're talking about.

Sounds like they're scanning all the most prominent repos and reporting results. Fine. But this is old code, a backlog that has never been looked at by LLMs before. Say those projects find your reports useful and patch then bugs. The backlog goes to near zero. What then? What are you going to use to impress your investors and sell your services on a regular basis?

Keep the layoffs as quiet as possible until it's go time to avoid panic. Announce on Friday at 4pm so that emoloyees don't have too much time to discuss between themselves and riot. Classic Office Space approach.

I can easily see how this central firmware delivery service will become the target of a malicious actor at some point. Just a matter of time.

It's despicable that it's taken 17 years for the courts to reach this verdict. It must have been a masterclass in corporate lawyer's delay tactics.

In the golden days of classic software engineering youd write tests of all sorts vefore the release - units, integration, smoke, performance, security... Then manual testers would come around and test the shite out of the system before going to production.

But now?

But how would you even begin to test nuggets like this? With all the human knowledge being compressed into a statistical model, how would even catch cases like this? You can't realistically hope to test against an infinite number of cases.

Dude. You're doing exactly what you shouldn't be doing. Your monologue above is nothing else than another attempt to tell me what I should eat.

Like I said - gonest what you want, but keep it to yourself. In other words - just shut up.

I'm not a snowflake.

I just get annoyed when other people tell me what I should and shouldn't eat.

Eat what you want but don't tell others what they should eat. It's like religion. Thy shall keep it to yourself.

"Non-meat mindset". You want to be a veggie? Absolutely fine. But shut the f up and stop telling other people what they should and should not eat themselves.

I weep for you. You've got a hard, misinformed life ahead of you.

For clarity, I meant scenarios where an attacker targets a process with root privileges, manages to execute code inside it and inheris root-level access, but then needs to load additional malicious modules for the second stage (lateral, persistence, etc.)

Removing unnecessary modules and reducing attack surface is great, but an attacker with root access could bypass this fairly easily. What you really need is Kernel Module Signing, so that you can't sideload malicious modules without a proper signature.

https://www.kernel.org/doc/htm...

It's no longer 1600s. We shouldn't be ditching information and context-rich language constructs just because some woke members of the society don't like it.