> In the case of GPT-Codex: it is on hold because the perceived relative impact for us is low

Interesting ... so for you "Hold" ~= "low impact / monitor ", while ThoughtWorks use "Hold" more like "don't start anything new" ( https://www.thoughtworks.com/en-us/insights/blog/technology-... ) [1]. Personally, I've started to read "Hold" in their sense - basically as "not recommended" for new work.

that’s why I first thought "GPT-Codex" (one of my favorite models) was already "not recommended." :-)

[1]

  ""
  Hold: The original intent of the hold ring was "proceed with caution", to   represent technologies that were too new to reasonably assess yet. But it has evolved into more of a "don't start anything new with this technology." You may be constrained to use it for existing projects because it is so deeply embedded into the tech portfolio, but you should think twice about using this technology for new development. 
  """

Thanks for the links - makes much more sense now!

EDIT:

Zalando definition ( https://opensource.zalando.com/tech-radar/ )

  "HOLD — Technologies not recommended to be used for new projects. Technologies that we think are not (yet) worth to (further) invest in. HOLD technologies should not be used for new projects, but usually can be continued for existing projects."

▲ leo_researchly a day ago | parent [-]

exactly. Maybe I need to update our ring-definition or least define them on the radar. Thanks!

BTW: what's your experience with GPT-Codex?

	▲	pella a day ago \| parent [-]
		Still in the honeymoon phase with gpt-codex `codex -m gpt-5-codex -c model_reasoning_effort="high"` it’s my current favorite, with claude-code as runner-up. As someone with aphantasia[1] I naturally lean toward backend / abstract modeling ( maps, technology radars, databases ) , so non-visual tools like codex-cli, claude-code, or even https://omarchy.org/ are especially appealing. I haven’t yet seen a technology radar that makes a visual vs. non-visual distinction. If you ever run a survey, adding such a category could be interesting - surprising patterns might emerge. [1] https://hn.algolia.com/?query=aphantasia