For this question, I would say big yess… couple of days ago I read in an article that Open AI is planning to launch their own email service to compete with gmail.
I’m not sure how authentic the news was but seeing their desperation to launch anything and everything, I won’t be surprised if they actually do so.
But with that they will have access to our emails, our conversation … Google already has in case of gmail but somewhere if i see from OpenAI perspective it looks bit scary.
There is less of scope for professional testers in these companies because in such companies there are researchers and engineers who do the testing of LLM based on certain parameters however there are company which are launching LLM based tools and in such testing skills can be used to evaluate the performance from different angles.
I’m curious to the level these work at in terms of prompts.
For example
Go to xyz.com and evaluate the site from an accessibility perspective, produce a brief report and recommendations to improve the site, store the script, actions and report and rerun weekly providing an additional comparison report for any changes regarding accessibility risk.
Is it this sort of target level, more or less?
I’ve clearly not done much digging into this, I remain wary it could be more simplistic and focused on doing things we have done in the past which perhaps should have been evolved into something better but just quicker and more efficiently.
I saw this video by @bethtestleadleeds. One of it’s kind and a quick demo. Nice and thanks.
A big problem that I see with the Open AI testing agent and other agents / no-code tools is that they all mistake testing as an activity about:
Pushing buttons
Navigating flows via GUI
Code (and how can they make it less coded)
Common things everyone can do (and doesn’t need any specialist)
Demonstration (that it works) rather than questioning the demonstration via actual exploration and experiments (that where does it not work).
Unfortunately, a lot of testing happening in the industry (I know mostly about the industry in India here) is about pushing screens & buttons. So, this kind of agentic stuff will certainly be used where testers are just doing that.
On the contrary, it will increase the demand for skilled testers / testing.
Also ton of questions. Lets say you become an agentic tester.
Is this a full time job? Suggestions seem to infer 10 mins a day.
What skills will you learn?
Who’s job are you aiming to replace?
Where is the fun and motivation of the role?
Will it improve stack optimisation of coverage?
On what sales basis pitch is it being sold? Cost and effort savings are red flags, non-skilled users are a red flag so curious as to the real value pitch beyond replacing and older less efficient practice.
