Joris de Jong

3.2K posts

Joris de Jong banner
Joris de Jong

Joris de Jong

@JorisTechTalk

doing AI stuff

Katılım Mayıs 2016
343 Takip Edilen3.6K Takipçiler
Joris de Jong
Joris de Jong@JorisTechTalk·
Billions of people will soon have access to creative AI directly inside the apps they already use. Meta is integrating Midjourney’s image and video generation into Facebook, Instagram, and Messenger. This could make AI image gen mainstream faster than any standalone tool could.
Alexandr Wang@alexandr_wang

1/ Today we’re proud to announce a partnership with @midjourney, to license their aesthetic technology for our future models and products, bringing beauty to billions.

English
0
0
2
421
agentnative
agentnative@agentnativedev·
finally cracked fullstack saas deployment on a plain old budget vps.. literally one command launch with hetzner, cloudfare, docker with lb and https (using traefik but few people also recommended caddy) next challenge is cicd via github actions and prometheus observability, then probably k8s + keda. im drafting a no fluff guide, drop a comment if you are in
agentnative tweet media
English
2
0
13
1.2K
Joris de Jong
Joris de Jong@JorisTechTalk·
Firebase Studio just launched. Three tries, three errors. Did anyone get it to work? @Firebase
Joris de Jong tweet media
English
3
1
3
705
Joris de Jong
Joris de Jong@JorisTechTalk·
Does anyone know if there’s a nicely formatted PDF of “On the Biology of a Large Language Model” by @AnthropicAI?
English
3
0
4
474
Joris de Jong
Joris de Jong@JorisTechTalk·
@yorgohoebeke That's awesome Yorgo. Any particular insights of the first two weeks which surprised you?
English
1
0
0
77
Yorgo Hoebeke
Yorgo Hoebeke@yorgohoebeke·
I launched my app 2 weeks ago - here's how it went... 🚀 155 new users 😐 Only 4 converted to weekly subscription 📈 75$ MRR Currently focusing on improving the app (web/desktop version + google calendar integration)
Yorgo Hoebeke tweet media
English
66
5
305
26.6K
Adam
Adam@_overment·
@DaveShapi I see; thank you for clarifying. > attention is the ability to keep track of large contexts. do you have any videos on this mb?
English
2
0
8
3.3K
David Shapiro (L/0)
David Shapiro (L/0)@DaveShapi·
umm guys Gemini 2.5 Pro just read 120,000 words in 20 seconds... and the attention seems fixed... guys... this might be game over
English
116
143
3.7K
421.1K
Joris de Jong
Joris de Jong@JorisTechTalk·
Gemini Flash 2.0’s photo editing capabilities blew me away. Built a product picture generator prototype in hours. Simple but functional frontend generated with @v0. Tech: NextJS, React, FastAPI, Gemini. It’s open source—see the GitHub repo below.
English
2
1
6
1.1K
Joris de Jong
Joris de Jong@JorisTechTalk·
Google’s new image editor, Gemini Flash 2.0 Experimental, is unreal. Built a quick prototype for a product image generator in just 1–2 hours. Imagine the products that are going to be built with this. I'll share the code later today.
English
2
0
10
1.5K
Joris de Jong
Joris de Jong@JorisTechTalk·
Want to add image generation/editing to your app? Gemini Flash 2.0 is still experimental, but so easy to implement with just a few lines of code. I built a simple prototype this morning, will share code later. Output isn’t perfect yet, but think of what it’ll do in 6 months.
Joris de Jong tweet media
English
1
0
5
544
Joris de Jong
Joris de Jong@JorisTechTalk·
Fundamentals > vibe coding No question. But, all these tools—@v0, @loveable, @boltdotnew —made prototyping nuts. Big respect to the devs busting their asses on v0, Loveable, Bolt.. Still takes skill to ship.
English
0
0
1
308
Jerry
Jerry@Ojie_jeremiah·
@JorisTechTalk @SokeyeA Hey Joris, cofounder of wetrocloud.com here. The best chunk size depends on relevancy and response time 1024 strikes the right balance from our research. Try wetrocloud, and we’d love to see a side-by-side comparison with your pipeline when it’s ready. Hope this helps!
English
1
0
1
91
Joris de Jong
Joris de Jong@JorisTechTalk·
Struggling with RAG chunk sizes in my vector DB, how many tokens max do you use for good semantic search?
English
1
0
4
604
Joris de Jong
Joris de Jong@JorisTechTalk·
@SokeyeA That's cool. Have you noticed any particular chunk sizes that work best? Balancing meaningful context without making them too large can be tricky, curious about what you’ve seen perform well.
English
2
0
0
38
Afolabi Sokeye 🧱
Afolabi Sokeye 🧱@SokeyeA·
@JorisTechTalk Oh ok I get that Wetrocloud only allows you choose the LLM for your RAG The entire RAG pipeline itself is end to end in the way it has been built so developers don’t have to worry about embeddings, chunking and vector databases.
English
1
0
2
54
Joris de Jong
Joris de Jong@JorisTechTalk·
@SokeyeA Haven’t checked them out yet. But I like building it myself, to understand what’s happening under the hood. Currently working with pgvector with Supabase integration. Can wetrocloud connect with database easily?
English
1
0
1
70

Keşfet

@_davideast @Firebase @AnthropicAI @yorgohoebeke @_overment @DaveShapi @sharqwy @v0