Greg_Ld (@Greg__LD) header
Greg_Ld (@Greg__LD) profile picture

Breton branché tech mais pas que,

Fes, Maroc - فاس، المغرب
Joined August 2022
Photo by Greg_Ld (@Greg__LD)Photo by Greg_Ld (@Greg__LD)Photo by Greg_Ld (@Greg__LD)Photo by Greg_Ld (@Greg__LD)Photo by Greg_Ld (@Greg__LD)Photo by Greg_Ld (@Greg__LD)Photo by Greg_Ld (@Greg__LD)Photo by Greg_Ld (@Greg__LD)Photo by Greg_Ld (@Greg__LD)Photo by Greg_Ld (@Greg__LD)
Update on Phonotek, my on-device take on ElevenReader, running locally on iPhone, built on Phonon by @GradiumAI: • now runs on Mac too • folders and favorites • app in English and French (Spanish, German and Portuguese coming soon) • nearly ready for the App Store Anyone at Gradium I could talk to?
1
62
I really like ElevenReader from @ElevenLabs. So when @GradiumAI released Phonon, I took the opportunity and ask Opus to build my own version overnight. It lives entirely on my iPhone: books, PDFs and articles read aloud, offline, 5 languages, 100+ voices. Just a side project, but I’m pretty happy with it.
4
1
1
13
1,938
Au chômage partiel
1
42
Replying to @AnthropicAI
@AnthropicAI please, please, please... do not... ever... nerf Opus 5.5
70
Job done, subscription ends in 5 days, ready for GPT 5.6 @thsottiaux
1
178
Thanks @mattshumer_ for finding one more gem use-case for the incredible GPT-Image-2. You can generate stop motion videos just by asking ! xitter.gq/mattshumer_/status/204…
GPT-Image-2 can be a video-gen model if you want it to be! Just ask ChatGPT to generate stop motion video, frame-by-frame. It works shockingly well and is super controllable!
4
3
25
23,028
please @OpenAI don't @AnthropicAI us and 5.5 in a few weeks from now !
1
87
Replying to @bcherny
@bcherny @AnthropicAI let's see where it goes from three weeks from now, but so far it has been the biggest disappointment. I was using Opus 4.6 with pretty much everything, but now 4.7 is just wasting my time with constant monitoring, stubborn refusal to perform tasks, loss of context... I don't see myself spending 100 or 200 a month for Sonnet... Hopefully you'll prove me wrong, but I'm now switching all agentic tasks to @NousResearch and coding with @OpenAI
46
Mes premiers tests de GPT-Image-2... j'ai lancé une recherche rapide de l'actu sur Grok et demandé à GPT de me faire une couv' façon Science & Vie... impressionant, j'attends juste les articles de @Frenchie_ et @ParadisLabs et je pourrais lancer l'impression lol
1
128
really dont know what to think about Opus 4.7 @AnthropicAI ... I really thought that all comments about it acting dumb were an organized campaign, but here I am with that crazy expensive model not able to run a playbook tht was flowing like a breeze with Opus 4.6 pre-possibly-nerfed version. Please @bcherny stop with the releases, concentrate on giving us back the product we used to happily pay $200 a month for
39
Replying to @oza_dev_
@oza_dev_ I was gladly surprised to come across Djeliaa while looking for online Bambara lesson, I have some feedback about the app, is it possible to get in touch with you or the team ?
1
1
1
1
309
I've discovered @badlogicgames through @nicopreme awesome page yesterday, this is the day my vision and use of agents completely changed. Pi is such a great project, I'm already itching to rebuild everything with it.
1
1
44
Mon bot Pinch a de l'humour, je lui ai posté ça, sa réponse m'a fait sourire
1
1
40
Non-dev here. Planned properly, gave Claude Code full access to my browser, let it run while I monitored every step. Result: complete webhook receiver (Fillout forms + Zoom bookings → GoHighLevel CRM) Fuzzy matching, Zoom sig verification, GHL v2 quirks… all handled. Live on Railway, processing real podcast guest onboardings. Feels like Make/n8n just lost a chunk of my future budget
2
102
This morning I saw @emollick post about making a quick explainer with Claude. And I thought—I want to learn something too. Formal logic. The foundations of reasoning. Problem: I can't learn theoretical stuff by reading. Never could. I need to do things. Test myself. So I decided to build one while tackling daily tasks. By 11am: logic-mastery-two.vercel.app 5 modules. 64 exercises. Interactive truth tables. Live. A year ago? This project dies in an error loop. I get stuck. Rage quit. Back to scrolling. That wall just... isn't there anymore. Not because AI writes code. Because I can brainstorm until the plan is tight, THEN build. Thinking first. Code second. When I get stuck—on tech or ideas—there's something to push through with. This is the window right now. You can experiment. Build your own tools. Learn how this stuff actually works. That's why @zarazhangrui latest article felt so important to me. The people messing around now get options. The people who wait? They'll use AI too. But it'll feel like catching up. Compliance instead of creation. I'd rather build than scroll. Healthier dopamine. And I learn something every time.
1
85
So I have an almost production-ready voice AI coach assistant that integrates with my interactive language lesson app, tracks your progress and topics of interest. The only issue so far has been the voice. Why am I only hearing about @GradiumAI and @kyutai_labs today? Félicitations à l'équipe pour Unmute ! Je vais me mettre au taff dès maintenant !
1
67
Replying to @gregisenberg
@gregisenberg hey you should tell you followers to report this guy : xitter.gq/gregisenbelg
7
Replying to @AnthropicAI
@AnthropicAI releasing features like this brighten my morning, day, and months to come...
7