Christian Laforte
@chrlaf
Followers
1,692
Following
416
Media
5
Statuses
110
Chief AI Scientist in a Stealth HW startup. Previously: Stability AI; AMD; Autodesk; CEO of 2 startups (3D + AI)
Toronto, Canada
Joined July 2023
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
Israel
• 4518146 Tweets
Iran
• 3105209 Tweets
ايران
• 1173898 Tweets
Presidenta
• 469251 Tweets
Middle East
• 397775 Tweets
Halloween
• 333796 Tweets
イスラエル
• 137820 Tweets
Wizkid
• 109166 Tweets
John Amos
• 105584 Tweets
Adams
• 95035 Tweets
Davido
• 93316 Tweets
60 Minutes
• 80951 Tweets
Margaret
• 59132 Tweets
Tampon Tim
• 52939 Tweets
#WWENXT
• 50512 Tweets
Moderators
• 48966 Tweets
Raiders
• 46246 Tweets
#VPDebate2024
• 42998 Tweets
からくりサーカス
• 40039 Tweets
ケンタッキー
• 33339 Tweets
Orioles
• 27662 Tweets
Frogido
• 27281 Tweets
Brewers
• 25500 Tweets
Nora
• 17621 Tweets
Giulia
• 14930 Tweets
Roxanne
• 14920 Tweets
FURIA Y LUCHI EN VIVO
• 14703 Tweets
Aces
• 14479 Tweets
#GandhiJayanti
• 14020 Tweets
महात्मा गांधी
• 12875 Tweets
Go JD
• 12686 Tweets
#マックのコーヒー記念カキコ
• 11242 Tweets
Gago
• 10097 Tweets
第三次世界大戦
• 10066 Tweets
Pinned Tweet
I left Stability AI in late July 2024 to join a new stealth HW startup. I want to thank my former Stability AI colleagues for a year and a half filled with exciting challenges, discoveries and technical achievements. I wish them all the best.
5
1
90
Hi, I was promoted to CTO at Stability AI last month, and we have exciting new AI products we'll launch soon. We're hiring 5 Full-stack Engineers () and a Senior Product Manager ().
28
27
234
@VikramVoleti
and I integrated and optimized a "2D⇒3D" feature in ThreeStudio -- an open-source text-to-3D and 2D-to-3D project. Here are recent results! (more results at the end of the thread)
2
23
139
@USEnglish215753
@StabilityAI
@EMostaque
Yes, the plan hasn't changed, we are still hard at work improving the model towards open release.
4
7
55
@rajdhakad_
@USEnglish215753
@StabilityAI
@EMostaque
Our plan is to soon release the API first to collect more human preference data and validate that our safety improvements don't cause the quality to suffer. Then we'll do some more fine-tuning (DPO/SFT) and release the weights and source code. Current ETA is 4-6 weeks.
9
2
35
@StabilityAI
I am privileged to be leading Stability AI (alongside co-interim CEO Shan Shan Wong) on its next phase of growth. Emad and the board named us both Co-Interim CEOs.
3
1
25
@Scobleizer
@StabilityAI
Thanks
@Scobleizer
. If your readers want an actual comment on the record of what it's like to work with me:
@Scobleizer
@StabilityAI
Yo, I enjoyed working with Christian very much. He is smart, focused and also a super nice guy.
1
3
31
0
0
21
@StabilityAI
I have deep admiration and gratitude for
@EMostaque
who has had an incredible impact on open AI. Emad, you have hired some incredibly talented researchers and engineers, that I will continue to nurture and steer towards impactful, profitable applications.
2
3
18
@StabilityAI
@EMostaque
Twitter folks: most of you don't know me. I don't plan on being very active here, but here's a brief self-intro. Emad hired me to start our 3D Generative AI Research 1 year ago, then entrusted me with increased responsibilities along the way: as VP of Research, then CTO.
1
1
15
@fakeryofficial
@thibaudz
Thanks on behalf of the team, however I don't personally deserve much of the credit, a lot of other people worked (and are still working) very hard on this.
0
0
15
@StabilityAI
@EMostaque
Prior to that I was a Technical Fellow at AMD and I also co-founded two startups that did 1) 3D Graphics and Computer Vision applied research and 2) mission-critical security software.
2
1
12
@OfficialLoganK
@StabilityAI
@JuliaLanguage
@EMostaque
Thanks a lot Logan! Best wishes on your next opportunity! 😊
0
0
9
@Imagine_aiart
Haha, good try, except you posted 1 day late for April's Fools. Stop claiming you have SD3, you don't. SD3 looks better than this.
4
0
10
1
0
7
@jampani_varun
Good work guys, I'll admit publicly I was skeptical this would work... I thought Stable Video Diffusion lacked the capacity to handle sudden large-scale disocclusions caused by significant changes in camera angle... I was pleasantly surprised this worked!
0
0
6
@StabilityAI
Congratulations to the research team (
@markb_boss
@jampani_varun
etc.) for this state of the art open release, and to
@ECourtemanche
for the cool animation!
0
0
6
@StabilityAI
I'm proud of our 3D Research team for this advance and we look forward to releasing this new capability to game and film studios, so they can produce better results more quickly and with less effort.
1
0
2
For these specific examples, we used a text⇒2D⇒3D process made fast and simple through
@clipdropapp
's integration of SDXL 0.9 (the world's best text-to-image model!) from
@StabilityAI
teams lead by
@robrombach
and
@MysteryGuitarM
1
1
4
Thanks
@StabilityAI
and
@EMostaque
for allowing us to work on this open-source
#GenerativeAI
3D project!
1
1
3
BTW our team has 2 new positions for ML+3D Research Scientists or Engineers... check my earlier post for the link
1
1
2
(projects that themselves build on breakthroughs such as
#StableDiffusion
and DreamFusion:
@poolio
,
@ajayj_
,
@jon_barron
,
@BenMildenhall
, etc. Thanks everyone!)
1
1
2
Our "
#GenerativeAI
for 3D and Games" team at
@StabilityAI
is hiring 2 research scientists or engineers. Please share widely!
1
1
2
@natanielruizg
@EMostaque
Hi Nathaniel, I'll reach out on Monday, hopefully we can arrange a quick meeting before I leave on vacations
1
0
1
For these specific examples, we used a text⇒2D⇒3D process made fast and simple through
@clipdropapp
's integration of SDXL 0.9 (the world's best text-to-image!) model from
@StabilityAI
teams lead by
@robrombach
and
@MysteryGuitarM
0
0
2
@StabilityAI
@tripoai
Well... I wish our marketing folks had picked a more interesting example video... follow the link for more examples.
0
2
2
This leverages the ThreeStudio open-source project (lead by
@superbennyguo
), improved code originally in Ashawkey Stable-DreamFusion, leverages the Zero123-XL model trained by
@ruoshi_liu
and
@mattdeitke
2
1
2
@vettukal
@EMostaque
Sorry, please try again soon, we're still trying to improve the autoscaling and how many A100s we provision... scaling will be one of the main focus the next few weeks
0
0
1
@TheZachMueller
It really depends whether a model generates a range of value (e.g. pixels or VAE latents) or it produces distinct labels/logits. If it's the latter, then it can be quantized down to ternary or int2/4 given enough effort. If it's the former, you need to sandwich quantized layers
0
0
1
- Generates a NeRF in ~5 mins on an A100 40GB, half of which is overhead (loading model etc) that can be optimized away relatively easily
- Improved generalization and robustness to the Janus effect (extra heads, wrong number of limbs) that plague many similar techniques
2
1
1
@PascalMassimino
@DaKangz
@GKopanas
Is it still working? I tried it on Chrome and Firefox (Ubuntu) and Chrome (Android) and they all produced this error:
1
0
0