@hakobtam

ML researcher @YerevaNN | ICML '25 spotlight | building in vision, VLMs, geospatial AI

 Armenia
Joined June 2020
  • Tweets34
  • Following379
  • Followers37
1Media
Not bad
Lucas Beyer (bl16)
@giffmanaAug 16
Guys:
5
427
When Mark posts it himself on X, you know it’s serious.
Mark Zuckerberg
@finkdJul 9
(1) Today we're releasing Muse Spark 1.1 -- a strong agentic and coding model at a very low price. It's available through our new Meta Model API and in Meta AI.
2
145
Async is setting a new value bar as TTS market is heating up.
Async Platform@async_platformJun 2
Big news for Async text-to-speech! Async Flash v1.5 just landed in the most attractive quadrant on Artificial Analysis’ Quality vs. Price benchmark with the best quality score among models under $10 per 1M characters 💪🏻🏆 That means Async now gives creators one of the strongest quality-to-price ratios in the TTS market, ahead of ElevenLabs, MiniMax, Microsoft, and OpenAI. High-quality AI voice is no longer a premium-only feature. It’s already inside Async. #AIVoice #VoiceAI #TextToSpeech #TTS #Async
1
3
167
#ICLR 2026 still 4 posters from our lab. Stop by if you're around
YerevaNN@YerevaNNApr 27
On the last day of ICLR workshops we have four posters.
122
Loving Codex @thsottiaux. Small bug though, rate limits drop just from browsing past chats, no model calls made.
68
Why is a standard feature like remote-SSH taking so long to release? Finally was able to use Codex app after those changes.
Hakob Tamazyan
@hakobtamApr 25
Replying to @chokt310055 @kimmonismus
Add this to ~/.codex/config.toml: [features] remote_connections = true Restart codex app, and then Settings > Connections. Still got several issues, like 5.5 wasn't available with remote ssh, but after removing codex cli from my remote, deleting and reinstalling the app on my Macbook, It works perfectly
1
177
Excited to see our YerevaNN team at @iclr_conf 2026 with 7 workshop posters.
YerevaNN@YerevaNNApr 21
We are happy to announce that six of our team members are traveling to Rio de Janeiro this week for @iclr_conf 2026. Our lab will be presenting seven posters across various workshops, highlighting our latest research in foundation models, molecular optimization, remote sensing.
180
Hakob Tamazyan retweeted
Chubby♨️
@kimmonismusApr 23
The accuracy gap in production voice AI is wider than most people realize. Async’s benchmark on non-standard text shows 81.2% sentence-level accuracy and 88.6% unit-level accuracy, tested on 2,200+ non-standard words in true real-time streaming. No preprocessing. No text rewriting. Link down below
🤝 Paid partnership
12
12
118
21,112
Hakob Tamazyan retweeted
Theo - t3.gg
@theoApr 13
Markdown is the new Python
238
273
60
4,817
318,713
Really hoping the next Codex app update adds this feature
Armen Aghajanyan
@ArmenAghaFeb 4
The Codex app is really nice, but I really wish it supported Remote-SSH. Most of my work requires GPUs on remote clusters/envs... adding this would make me move to Codex full-time.
2
108
Hakob Tamazyan retweeted
huss
@hussamfyi21 May 2025
323
4,730
644
31,467
2,113,988
Hakob Tamazyan retweeted
shweta
@shwedotai16 Apr 2025
With Al removing accents, it'll be impossible to tell if "Josh from Texas" is actually an Indian dude dialing in from a Bangalore call center. (accent neutralizing starts around 0:22)
492
564
587
6,300
1,691,306
Excited to present our new benchmarks for remote sensing foundation models at @eccvconf 2024 in Milan during the OOD-CV poster session on September 30!
YerevaNN@YerevaNN27 Sep 2024
YerevaNN's Hakob Tamazyan is set to represent us at @eccvconf 2024 in Milan. He will present our new benchmarks to measure generalization of remote sensing foundation models at the OOD-CV workshop. Visit our poster session on September 30!
5
101
2/ UForm Pocket sized multimodal AI for content understanding and generation. Impressive results, and 5x smaller than LLaVA and BLIP @ashvardanian @unum_cloud
1
14
3
279
201,872
Hakob Tamazyan retweeted
Armen Aghajanyan
@ArmenAgha14 Jul 2023
I’m excited to release our most recent work setting a new SOTA FID of 4.88 on text-to-image generation we call CM3Leon (pronounced chameleon)! ai.meta.com/research/publica…
24
121
17
459
166,245
Load more