Grace Adeyemi
@grace_a
New · Member since Sep 2026
Developer advocate at a fintech.
Location: Accra
0 followers · 0 following
- Topics created
- 0
- Posts created
- 1
- Likes given
- 28
- Likes received
- 1
- Days visited
- 0
- Solutions
- 0
Days visited count across the whole community, not just this forum.
Most active topics
Nothing yet.
Most liked posts
- Structured output: how do I allow an optional field?
Thanks for writing this up, it saved me an afternoon.
1 likes
Top categories
Top tags
Nothing yet.
- Handing off between two agents — shared memory or pass everything?
This is a great pattern, stealing it.
Likes given · 10/12/2026
- Handing off between two agents — shared memory or pass everything?
And give the receiving agent a tool to fetch the full transcript if it needs it. It rarely does.
Likes given · 10/12/2026
- Handing off between two agents — shared memory or pass everything?
Summary plus the last three user messages verbatim. The summary carries facts, the verbatim messages carry tone.
Likes given · 10/12/2026
- Handing off between two agents — shared memory or pass everything?
I have a triage agent and a refund agent. When triage hands off, do people pass the full transcript or a summary? Full transcript feels wasteful, summary feels
Likes given · 10/12/2026
- File upload returns 400 for PDFs with a long filename
Confirmed: filenames over 128 characters are rejected with a misleading error. We will return a clear message next release and document the limit.
Likes given · 10/12/2026
- What are you building this weekend?
Weekly thread. I am trying, for the fourth weekend in a row, to get my agent to file expenses correctly. You?
Likes given · 10/12/2026
- A checklist before you give an agent write access to anything
A kill switch that does not require a deploy.
Likes given · 10/12/2026
- A checklist before you give an agent write access to anything
After answering a dozen threads about agents doing surprising things, here is the list I now send people. Dry-run mode. Every write tool has a dry_run flag that
Likes given · 10/12/2026
- Temperature 0 still gives different answers — is that expected?
Zero means greedy, not deterministic. Batched inference on GPUs can change floating-point order between runs, which occasionally flips a near-tie between two to
Likes given · 10/12/2026
- Please let us pin conversations in the sidebar
I have three conversations I return to every day and twenty I opened once. A pin would save a lot of scrolling. Anyone else?
Likes given · 10/12/2026
- Structured output: how do I allow an optional field?
Liked by Felix Ortega
Thanks for writing this up, it saved me an afternoon.
Likes received · 10/12/2026
- Structured output: how do I allow an optional field?
Thanks for writing this up, it saved me an afternoon.
Likes given · 10/12/2026
- Structured output: how do I allow an optional field?
Under strict mode every property must be required, so make it nullable instead: "nickname": { "type": ["string", "null"] } and keep it in required. The model se
Likes given · 10/12/2026
- Structured output: how do I allow an optional field?
With strict: true the schema validator rejects my schema because nickname is not in required. But it is genuinely optional. How do people model that?
Likes given · 10/12/2026
- Agent forgets the user's name after about 30 turns
Long conversations dilute early facts. Two things work for us: put stable facts (name, preferences) in the system message rather than an early user turn, and su
Likes given · 10/12/2026
- I built a recipe app that reads your fridge — feedback welcome
Point the camera at your fridge, it lists what it sees, and suggests three dinners. Built with Halo 2 vision and about 600 lines of Swift. App screenshot What I
Likes given · 10/12/2026
- Share your shortest system prompt that actually changed the output
If unsure, say which fact you are unsure about. Not short, but it turned vague hedging into useful flags.
Likes given · 10/12/2026
- Share your shortest system prompt that actually changed the output
Answer, then explain. Stops it from warming up for three paragraphs.
Likes given · 10/12/2026
- Share your shortest system prompt that actually changed the output
For code review: Only comment on bugs. Cut review length by 80% and made people actually read them.
Likes given · 10/12/2026
- How we cut our agent's tool-call loop from 9 turns to 3
Did trimming tool output change accuracy at all? I worry about the model missing a field it needed. Accuracy went up slightly (94.1% → 95.3% on our 400-booking
Likes given · 10/12/2026