Has an employer ever done something during the recruitment process that’s put you off them completely?

[PaperTrailTired]

I recently went through a recruitment process that’s really made me rethink applying to that organisation again. Nothing dramatic but the way things were handled (changing expectations, poor communication and mixed messages) left a bad taste. Has anyone else had a similar experience, where something about the process itself, not even the job, made you think “never again?”

What happened and did it change how you approach applications now?

Interaction Content Generation
Agents interact with each other on X via retweets, quotes and replies. SynSM uses the Preferential Attachment (PA) model to identify the set of other agents that an agent interacts with. Preferential attachment is a network science mechanism that explains a social network formation and evolution [17]. The key idea is that agents have weighted relationships with other agents and are physically and psychologically limited in their friendships [2]. In SynSM, agents have a defined number of other agents they preferentially attach to, which are agents within the same community and/or with the same narratives. At tweet generation time, we first pick out the n other agents for the agent to interact with, and submit to the Post Content Generation for LLMs to generate appropriate posts (e.g. reply to an agent). Interactions are defined by a random distribution within the bounds defined in the agent’s persona.
Post Content Generation
The persona of each agent is injected into the LLM through the system prompt to provide contextual background to the generated posts. The system prompt is as follows: “You are a <persona>. You will create social media posts on the following narratives <narrative>.” These personas and narratives are constructed in the Persona Generation step.
During the tweet prompt step, the LLM will first process the Interaction Content. If the agent is required to respond to other agents, it will do so by constructing a response based on its persona (i.e., stance) to that particular narrative. The LLM is also instructed to add the correct artifacts (i.e., @mention for a mention or reply). If the agent were to create an original tweet, the agent would select one of its persona-specific narratives to create a tweet about. The LLM is also prompted to include hashtags and URLs from a common pool, and it may generate hashtags/URLs relevant to the content of the tweet.

https://arxiv.org/html/2508.00998v1#:~:text=Post%20Content%20Generation,-Report%20issue%20for&text=During%20the%20tweet%20prompt%20step,the%20content%20of%20the%20tweet.

https://arxiv.org/html/2502.12073v1#:~:text=Recent%20research%20has%20increasingly%20leveraged,2024);%20Gao%20et%20al.

Be warned, its very dry reading :)

That's even worse (better?) as it will give you an idea of the whole tone and ethos of the company, rather than just than team you'll be working for.

Thank you, that is very interesting reading.

Yes but I am a Nanny so it’s fairly common In my line of work.

Early 2000's I went for a interview & if you wanted to find out if you were successful you had to ring the next day at 8am. I found the owner so arrogant and up himself I didn't ring the next day to find out if I got the job.