axel padilla

440 posts

axel padilla banner
axel padilla

axel padilla

@adpadillar

swe + cofounder @ https://t.co/sjGWKXoVQs

guadalajara, jalisco Katılım Nisan 2022
313 Takip Edilen38 Takipçiler
axel padilla
axel padilla@adpadillar·
@DanniellaMariex @jrdeuf @F1_Papi I agree George has been unlucky and has had more issues outside of his control, but I don't think you can deny that Kimi just literally matched Lewis Hamilton best winning streak (5 in a row) the first time he actually has a dominant car. Def circumstantial, but still impressive.
English
1
0
1
177
Danni
Danni@DanniellaMariex·
@adpadillar @jrdeuf @F1_Papi George has had issues with the car in nearly every single race - it literally DNF’d itself while he was in the lead and Kimi was making mistakes every 5 seconds. So yes, his machine is superior to George’s but on reliability only.
English
1
0
5
145
F1_Papi
F1_Papi@F1_Papi·
En 5 meses, Antonelli va a superar lo que Leclerc ha logrado en casi 8 años. En 3 meses va a superar todo lo que Carlos Sainz ha hecho en su carrera de F1. En la próxima carrera igualaría a su compañero, George Russell, y a Sergio Perez.
Español
47
346
8.8K
218.8K
Javier Rodríguez
Javier Rodríguez@jrdeuf·
@F1_Papi Ese tipo de estadísticas son las que hacen de la F1 un deporte difícil de entender en mi opinión. No significa que sea mejor que cualquiera, solo dice que tiene una máquina superior.
Español
4
0
83
6.3K
Jay
Jay@jayair·
It’s crazy, they won’t stop talking about it I think it’s the placebo effect tbh
English
8
0
83
10.1K
axel padilla
axel padilla@adpadillar·
@charlierguo @timurkhakhalev Can we maybe then reduce the usage 5.4 takes in subs? This is the main reason a lot of us are still in codex, and I get that compute is limited but raising the usage users get of 5.4 even if it doesn't match 5.3 codex would be very good faith of you
English
0
0
0
176
Charlie Guo
Charlie Guo@charlierguo·
@timurkhakhalev Starting with 5.4, we merged the -codex models into the mainline models! All of the coding capabilities are still available in 5.4 and 5.5
English
3
0
29
1.6K
Timur Khakhalev
Timur Khakhalev@timurkhakhalev·
Nooooooo!! Not 5.3-codex!! We don't have any other same model as -codex models! I use it in my workflows to write a code following the plan and this works great. We need new -codex model bro
Tibo@thsottiaux

To simplify our Codex compute fleet management, we will be sunsetting GPT-5.2 and GPT-5.3-Codex in Codex on June 2nd when logged in with your ChatGPT account. For free plans, GPT-5.5 will be the default frontier model to build and work with going forward. These models will remain available on our API.

English
27
7
340
57.8K
axel padilla
axel padilla@adpadillar·
@thsottiaux Nooo pls I still use 5.3 codex everyday. 5.4 is slightly better but goes through usage faster
English
0
0
0
28
Tibo
Tibo@thsottiaux·
To simplify our Codex compute fleet management, we will be sunsetting GPT-5.2 and GPT-5.3-Codex in Codex on June 2nd when logged in with your ChatGPT account. For free plans, GPT-5.5 will be the default frontier model to build and work with going forward. These models will remain available on our API.
English
637
152
4.6K
778.7K
axel padilla
axel padilla@adpadillar·
@theo i feel like the people themselves are not the problem. it is something else
English
0
0
0
183
axel padilla retweetledi
Logan Kilpatrick
Logan Kilpatrick@OfficialLoganK·
Gemini
Indonesia
272
158
3.2K
525.8K
axel padilla
axel padilla@adpadillar·
@thdxr I actually assumed this is how it worked except for desktop app, which seems to sidecar spawn its own backend
English
0
0
0
96
dax
dax@thdxr·
one pattern we could do is the first time you run opencode it forks into the background as a server then every other time you launch it or use the webapp or desktop app they all use that one instance so everything is synced i'm worried this feels unexpected to people though
English
116
3
204
39.5K
Theo - t3.gg
Theo - t3.gg@theo·
Rebuilding the T3 Code marketing site. Want to add nice quotes from people about T3 Code. Reply with some below and I'll put my favorites on the site :)
English
344
6
715
84.7K
Jared Palmer
Jared Palmer@jaredpalmer·
Excited to share that I’m joining @Xbox as VP, Engineering & Technical Advisor to CEO @asha_shar I’ll be focused on building world-class tools, services, and experiences for developers and players across the Xbox ecosystem. Grateful for the opportunity and excited to get to work.
Jared Palmer tweet media
English
475
460
8.8K
1.1M
axel padilla
axel padilla@adpadillar·
@RoyAllenFSW @AppleTV @F1 A weekend you'll get great tv direction only for the next weekend to miss all the important things at the right moments. This weekend felt particularly bad
English
0
0
0
7
axel padilla
axel padilla@adpadillar·
@RoyAllenFSW @AppleTV @F1 I agree, you should complain, but to F1, not necessarily apple. World feed is the same for everyone, and sometimes it's fine, but more than it should its just bad. It is notoriously weekend dependent, it seems like it is not the same people week in and week out
English
1
0
0
8
Roy Allen
Roy Allen@RoyAllenFSW·
This @AppleTV @F1 broadcast is almost unwatchable. 4 time World Champion does a 360 in the opening lap and it's not replayed until the 8th lap? A car rolls over and the commentators are guessing what happened when there are multiple cameras on every car?
English
145
25
2K
123.3K
axel padilla
axel padilla@adpadillar·
@peach2k2 I run out of my session in like 3-4 prompts in the 20/mo plan. I simply ask for reading a linear issue and implementing it on codex 5.3 and it goes for like 20 minutes and that's enough for like 20-25%
English
0
0
1
63
« 2k2 »
« 2k2 »@peach2k2·
ts feels nigh endless, how do you guys even run out??? just the $20 tier btw
« 2k2 » tweet media
English
26
0
147
16.2K
† lucia scarlet 🩸
† lucia scarlet 🩸@luciascarlet·
@xlatentspace pointless to waste tokens on very minor changes I can easily make myself (especially if it could get them wrong) I’m broke and have Plus so my tokens are very precious to me 🥺
English
2
0
13
782
Julius
Julius@jullerino·
New model picker live in latest nightly! ⌘+Shift+M or /model to open ⌘+1..9 to quick select, or navigate with arrow keys 🔍Search ⭐ Favorite Thanks chrono for the initial work copying @r_marked's design from @t3dotchat
English
42
6
506
33.6K
axel padilla
axel padilla@adpadillar·
@jullerino I have codex 5.3 trying to figure it out, will post if it finds anything useful that is easy enough to fix
English
0
0
0
54
Julius
Julius@jullerino·
@adpadillar Will look into it! I thought we dynamically picked up agents and variants but looks like thats not the case. Will fix!
English
1
0
1
334
Julius
Julius@jullerino·
OpenCode and Cursor (early access) support is now in latest Nightly builds. Try it out, send any feedback. Hoping to promote to latest soon!
Julius tweet media
English
58
8
615
87.3K
axel padilla
axel padilla@adpadillar·
@jullerino also not every model that has configurable thinking effort can currently be configured. if i use Codex 5.3 via Opencode OpenAI provider then i can configure effort, but if i use the same model via Github Copilot provider then it is not configurable
English
0
0
0
48
axel padilla
axel padilla@adpadillar·
@jullerino or also be able to configure the models that are surfaced per provider in the configuration, instead of showing all of the available models all the time
English
1
0
0
24
Julius
Julius@jullerino·
@Hazihell But u can use your copilot sub through opencode as well
English
2
0
2
209
axel padilla
axel padilla@adpadillar·
@OwenCrossssss @jullerino i mean there is always `opencode upgrade` command, t3code could call that on startup or periodically to keep it up to date i think
English
0
0
1
39
Owen
Owen@OwenCrossssss·
@jullerino Weird question, does the opencode sdk provide anyway for t3code to check it for updates and maybe kick off updates for it?
English
1
0
2
614

Keşfet

@DanniellaMariex @jrdeuf @F1_Papi @jayair @charlierguo @timurkhakhalev @aphumphreys @thsottiaux