RE: https://mastodon.social/@bagder/116639356762841470
Can't help but think about all these people who are busy finding vulnerabilities in existing projects, the companies behind them, the hundreds of billions that subsidized LLM development and the hardware running them without which this extreme fuzzing wouldn't have been possible.
And literally no one moving to the other side of the fence, rolling up their sleeves and actually fixing the bugs.
@gabrielesvelto when the point of these LLM security tools is to generate headlines and awareness of their products, dumping the CVE on the world and not writing a patch is a feature, not a bug. So many eyes on their product when every tech news site is reporting on the days of panicked patch writing and patching.
> And literally no one moving to the other side of the fence, rolling up their sleeves and actually fixing the bugs.
Er... what... I had a batch of these in my FOSS project in March and April and they were all fixed within a few hours.
Bugs and maintainer jobs existed before AI and they'll exist afterwards, made easier if the maintainer uses AI to solve them. AI will also "solve" coordination being a problem, because there will be no window to coordinate.
@hopeless the role of a maintainer in most FOSS projects is not a job, it's volunteering. It's absolutely not guarantee to exist in the future unless the industry steps up the funding, which it is not. That being said you seem to be working on the assumption that "AI" is going to stick around which is extremely unlikely at this stage.
@gabrielesvelto Yes same here... nobody pays me and I don't want / have a job for years now.
> That being said you seem to be working on the assumption that "AI" is going to stick around which is extremely unlikely at this stage.
I know it feels like that if you believe what you read here (and ignore what is written everywhere else). However the ignored key point is... coding AI works extremely well. It's not going anywhere, no matter if all the datacentres get droned, it will run locally.
@hopeless it's not about what you read here, it's about the largest companies in the tech sector having erased their cash-flow or gone cash-flow negative while loading themselves with 100Bs $ of debt. Who's going to spend the billions required to train a model if it's a losing proposition? What revenue does a local model provide? And that's ignoring the problem of model collapse which is only kicking into gear now, as the training data gets poisoned with LLM outputs.
@gabrielesvelto Fixing bugs doesn’t enter the news cycle ¯\_(ツ)_/¯
@bagder
@gabrielesvelto Ed Zitron, in his latest newsletter, says “The AI bubble is entirely built by people who hope somebody else will solve their problems.”
Install bonfire.cafe
Get the full app experience
