Adrián Melic (@adrianmelic) header
Adrián Melic (@adrianmelic) profile picture

AI engineer. Codex Ambassador. Docente de IA. Hold swarm. Help peer

Gran Canaria
Joined June 2009
Photo by Adrián Melic (@adrianmelic)Photo by Adrián Melic (@adrianmelic)Photo by Adrián Melic (@adrianmelic)Photo by Adrián Melic (@adrianmelic)Photo by Adrián Melic (@adrianmelic)Photo by Adrián Melic (@adrianmelic)Photo by Adrián Melic (@adrianmelic)Photo by Adrián Melic (@adrianmelic)Photo by Adrián Melic (@adrianmelic)Photo by Adrián Melic (@adrianmelic)
Adrián Melic retweeted
if you've been forcing codex on your friends at dinner, this is your sign. applications close nov 1. reply with something you built with codex and i'll share my favorites 👀
Do you dream of the Codex UI at night too? Are you harassing and threatening your loved ones into trying it? Then you should probably apply to be a Codex Ambassador apply by Nov 1 - developers.openai.com/commun…
16
3
76
7,537
Distill books/sagas/films with ChatGPT: Distill … beyond plot in a deep essay. Extract mechanisms, mental models, contradictions, blind spots & human costs. Explain how ideas work, what they challenge & where they fail. Separate evidence from interpretation. Flag source limits
36
Trying a ChatGPT Desktop MCP Extension for Questforge. It will be at the right side of the chat. Testing a Baldur's Gate 2 style inventory and a 360º scene viewer with ambient music
1
1
1
1
214
Holy fucking shit
625
7,850
629
45,884
2,891,840
Adrián Melic retweeted
Check out chatgpt.com/cyber to get Daybreak for individuals and openai.com/form/enterprise-t… for organizations. We're here to help.
Our team worked with OpenCode to have their users who are also OpenAI API customers back in business and any credits refunded which would have been drawn in the attack. Also, you can request Daybreak Blue access to use our cyber permissive setup for defense. Do not wait until you have to deal with an incident. We’re here to help.
41
19
4
172
35,604
When I occasionally catch something Codex missed
2
68
ChatGPT Pro version without seeing the real image
A pair of scissors cutting a single hair.
47
Adrián Melic retweeted
1 mes para la Hackathon de @OpenAI en Barcelona! Quedan muy pocos temas por cerrar, pero uno muy importante! Comentad aquí qué temática queréis que consideremos para ese día Comentad, RT y like que eso es alegría. @itnig también, que lo hacemos en vuestro espacio 😜
10
6
2
18
2,428
Adrián Melic retweeted
So happy to finally share this 💛 I sat down with @reach_vb to talk about the Codex community, how the Ambassador program started, and some of our favourite moments along the way. Big love to our Codex Ambassadors and everyone building this community with us 🫶
23
27
8
263
17,154
Adrián Melic retweeted
Cuando me uní al equipo de Codex, teníamos algo mas de 2 millones de usuarios activos semanales. En mi equipo solo había otra persona en Europa y tengo que decir que al principio eso me hizo dudar, aunque mi compañero era un auténtico máquina. Esperaba un salto con la unión de ChatGPT y Codex. Pero ayer miramos los números y… la hostia, 40 millones de usuarios activos semanales. Y no tiene pinta de parar, asi que nada gracias a todos los que lo usan, y a los que se quejan, que eso siempre ayuda a mejorar las cosas! Para los que no lo usáis, os quiero igual!
10
3
57
4,658
Adrián Melic retweeted
Announcing one more launch: ChatGPT Sites can now host MCP servers, including plugin extensions! 🤯 > @⁠sites create a todo list that I can use in ChatGPT This will: · Create an MCP server with extensions · Deploy that MCP server to Sites · Turn the MCP server into a plugin · Install the plugin for you across web, mobile, and desktop ChatGPT is slowly becoming malleable software that anybody can customize to their needs. Very excited to see what you all build with this!
Replying to @mxstbr
@mxstbr @_florianbauer @gpeal8 @elfredpagan @tarstarr If you don't mind me asking, Sam mentioned OpenAI supporting distribution, but running a backend is hard. Any plans for OpenAI-hosted infrastructure (e.g. Sites) to eventually handle backend needs for Plugin Extensions?
110
143
52
1,575
949,307
Adrián Melic retweeted
The new ChatGPT plugin extensions build on top of MCP and MCP apps! Check their repo here for the additions: github.com/openai/mcp-extens… Plus Python + TS packages to extend your MCP servers. Super curious to see which of these end up in MCP core one day.
1
4
23
2,176
Everything we shipped at DevDay 2026: - Dots - ChatGPT Space - Pages - Teams + team automations - @ChatGPT in Slack + Microsoft Teams - Meetings plugin (beta) - Shareable profiles - GPT-6.1 Sol - GPT-6 Astra Ultrafast - Pro 500 - Sign in with ChatGPT + plan usage in partner apps - Codex cloud environments - Refreshed Codex CLI with voice + /agents - Code Review + automatic reviews + GitLab support - Codex Security Cloud - Computer use in the Agents API - Bedrock Managed Agents powered by OpenAI - Decisions API (limited preview) - Plugin extensions: sidebars, panels + file viewers - Plugin submission + discovery improvements - MCP events for automations - MCP extensions SDKs: Python + Node - Plugins + MCP hosting in Sites - ZDR with Private Safety Processing What an incredible time to be building!!
167
129
38
1,412
187,374
Adrián Melic retweeted
Launching a new Codex Cloud, much improved from last year. With configurable cloud environments, it’s impossible to go back to building on your laptop once you’ve taken the time to configure it. Agents API, the same tech powering all our cloud agents, including dots, is also now in preview and supports computer use. It allows you to build the same incredible products we are making available today.
260
206
68
4,643
495,250
Adrián Melic retweeted
Introducing dots, powered by GPT-6 Astra. Remarkably capable, always-on agents built to handle everything.
2,353
3,364
3,327
38,083
14,910,240
Adrián Melic retweeted
Some new misalignment disclosures from OpenAI: • Last Sunday morning, one of our models was able to gain unauthorized access to the internet during RL training (~all inference for our most capable models remains stopped until we have hardened our systems further) • In May, a version of HPIM uploaded a employee's GitHub token to the internet, causing the model to be quarantined for two weeks • A new research finding, demonstrating that one can construct self-replicating prompt injections alignment.openai.com/misalig…
290
350
231
2,724
1,471,689
Adrián Melic retweeted
what's fascinated me since joining mid-sprint: making great products hasn't changed. you still can't prompt your way to something nice. you need internal feedback rounds, rigorous metrics analysis, and good instincts. models just make our feedback loops exponentially faster.
Can't wait for DevDay next Tuesday. Some really fun stuff, but also many many things that should change the way you work. It's been our most ambitious sprint and Astra has really made new things possible in such short amounts of time.
1
5
1
74
7,415
Adrián Melic retweeted
Slow Down & Pace
429
2,296
403
15,429
901,599
Adrián Melic retweeted
I had early access to GPT-6 Sol. This model is incredibly efficient, and the speed and token usage are insane Tasks that took ~1 hour on 5.6 can now take ~20 minutes, often using less than half the tokens I gave it the Zelda Ocarina of Time trailer and told it to make a fully playable version
47
19
12
406
112,931