![](/static/253f0d9b/assets/icons/icon-96x96.png)
![](https://feddit.de/pictrs/image/9067d8fe-73da-4969-8f64-4b9ba7b3105e.png)
Hier gibts einen Artikel zum personalisierten Spam:
https://www.bcg.com/publications/2023/gpt-was-only-the-beginning-autonomous-agents-are-coming
For instance, in the near future, an autonomous agent could allow a marketing executive to carve out and automate whole segments of work. Based on a company’s past marketing campaigns, the agent could determine what worked and what didn’t, making its own decisions for future email design, scheduling, graphics, and subject lines. It could also identify the types of consumers a campaign should target and then assess whether the results—opens, views, clicks, and responses—are worth reporting back to management. If the results fail to meet the campaign’s objective, the agent could independently start again, creating a new, more refined list of target customers based on responses to the previous campaign.
Usually a sign of multiprocessing/multithreading going wrong, e.g. accessing the same resource without proper locks like opening the same logfile in different processes and trying to write simultaneously. Those errors can be triggered just by reformating the code (or obfuscating in this case), thus changing the runtime behaviour slightly. Hard to find, especially since they’re dependent on the speed/workload of the machine running the code.