Mogtaba Alim @MogtabaAlim
Giving Every Model a Voice Prev @Amazon AGI, @awscloud, @UHN University of Toronto Joined October 2024-
Tweets117
-
Followers129
-
Following300
-
Likes486
Such a great story!! 👏 This line really caught my eye: “From the inside, the market didn’t look finished. It looked like it had barely started.” That’s exactly how voice AI feels right now. Still so early, so much left to build. What a time to be building in voice 🔥
In just 1 year, Fish Audio reached $21M ARR, 8M users, and raised a $52M seed. When we started out every investor and friend told me it was too late to compete in voice AI. Turns out, it wasn't. Here's the story of our first year.
So much credibility is lost when folks say 5-10 years to cure all human disease. We can't even measure most of human disease!
I'm keen on AI having a transformational impact on human health. But these cure projections are wildly off-base and impossible.
@ziyadbasheer @amrmkayid Fire launch 🔥🔥 Congrats!
I'm excited to share what we're been on working on at Aide: making AI agents reliable for production use in customer-facing interactions. 1/ 🧵
Speech-to-speech no longer needs speech-to-text! Until now, our stack was VAD -> STT -> LLM -> TTS. Now it can send audio directly to multimodal LLMs: VAD → MLLM → TTS No STT. The model understands your voice. Now go build better voice agents! github.com/huggingface/sp…
@andimarafioti When will we be able to replace VAD too?
I have several thoughts after watching this, but the big takeaway is that this is being framed wrong. This was an alignment failure first and a security issue second. And yet, the primary take home message they prioritized was basically “buy our product and to defend yourself.”
Blackhat has uploaded the full presentation on the OpenAI Hugging Face incident, about which much ink has been spilled. youtu.be/87DyyMV0kCY?si…
Today, “full-duplex” is one of the hottest terms in voice AI. Funny enough, when we released Moshi, the first and until recently only full-duplex model, I didn’t even know that term existed, or that turn-taking was considered a fundamental issue. We were simply trying to build a conversational model. We had no prior experience with dialogue systems, so this end-to-end approach felt like the most natural one. I only learned the term after @JulianSlzr tweeted about our work, and realized that we had made a bigger contribution than we initially thought.
How far are we from "Her"? I couldn't help but wonder, given the rapid fire of GPT-Live and Thinking Machines. Turns out, full duplex speech-to-speech is not particularly new. Moshi has been around for more than 2 years. Many thanks to @neilzegh for bringing so much nuance.
Canadian led seed round
We created a pitch deck to tell a handful of VC firms about us and what we were up to (a fun experience!). Here’s a few slides about our background and some of the things we’ve worked on from the pitch deck (it was fun putting together the list of people in our teams who have
Announcing Discovery Loop! I am very excited to announce that, along with my longtime friends and collaborators @Sanjay_Ghemawat, @OriolVinyalsML and @quocleix, we are founding Discovery Loop (@DiscoLoopAI), a Public Benefit Corporation whose mission is to automate machine learning, science, and engineering to accelerate discoveries and progress. The four of us have worked together for 14 to 30 years, and have helped build some of the world’s most used products, infrastructure and AI models, and we’re excited to turn our attention to this ambitious endeavor. ♾ Learn more at: discoveryloop.com
you know RL doesn’t really generalize even at frontier scale bc lab employees are always like “let us know what the model isn’t good at and we’ll fix it for the next one”
Introducing Kimi K3: Open Frontier Intelligence 🔹 2.8 Trillion Parameters, 1 Million Context, Native Multimodal 🔹 Kimi Delta Attention enables up to 6.3x faster decoding in million-token contexts 🔹 Attention Residuals deliver ~25% higher training efficiency at <2% additional cost 🔹 Built for long-horizon agentic coding and self-evolving workflows Kimi K3 is now live on on Kimi.com, Kimi Work, Kimi Code, and the Kimi API. Open Weights by July 27, 2026. 🔗 API: platform.kimi.ai 🔗 Tech blog: kimi.com/blog/kimi-k3
.@UofT has announced a new, multi-year partnership with @cohere, which was founded in 2019 by former @UofTCompSci students and has become the world's leading sovereign AI company. utoronto.ca/news/u-t-partn…
The university of Toronto was such a massively positive influence in my life. It’s where I met my cofounders, it’s where I met my band, and it shaped so much of what I think about today. I’m so thrilled to be working on this with them now :)
.@UofT has announced a new, multi-year partnership with @cohere, which was founded in 2019 by former @UofTCompSci students and has become the world's leading sovereign AI company. utoronto.ca/news/u-t-partn…
Great to finally see a true "Open" AI release
Today, we are introducing Inkling. Inkling reasons efficiently across text, image, and audio modalities. We are making the full weights available. thinkingmachines.ai/news/introduci… Available today for fine-tuning on Tinker. Play with it in the Inkling Playground. 🧵
We’re committing $10 million CAD and partnering with leading AI institutions in Canada to help fund new AI research. anthropic.com/news/canadian-…
Open source harness you say
It’d be great if model releases came with a one-pager on their data focus. Im not even expecting a full technical report atp GPT 5.6 feels unusually well-suited to voice orchestration. During testing when I correct a mistranscribed word it often starts its response with “Sorry I misheard you” even though it only received a text input
"The first step of any project is to grossly underestimate its complexity and difficulty." — Nicoll Hunt
Such a great resource for people learning about or building realtime voice AI! 🔥 @lvwerra @Thom_Wolf wondering about the technical decisions using WebSockets over WebRTC or using something like @livekit 🤔
Most people should probably update their priors on the state of open-source speech-to-speech. It's honestly kind of mind-blowing. We teamed up with @cerebras to build a fully open-source realtime voice demo (models + code) to show what's possible today. Demo :
⊹ maddie 💌 🎀 @5455ym4dd13x
3 Followers 144 Following hewwo im maddie 💌 here to simp and be simped on 😇 followback always
Ka @its6i77
2K Followers 558 Following lead engineer @HUMAIN | اللهم ارحم أبي | Majoring Cs at @uwaterloo
alvvin @alvanliii
453 Followers 3K Following this account is for shitposting, my other acc @alvanlii
Daryl W @Daryl_wanji
286 Followers 230 Following Engineer brain. Accidentally fell into mechanistic interpretability. Speak in weights & bias
Jim Banting @JimDBanting
1K Followers 2K Following Supporting partnerships, innovation, and entrepreneurship. Past: biotech bus dev (🇺🇸 & 🇨🇦) & co-founder. Views are my own.
VC Intern @the_vc_intern
2K Followers 3K Following Tracking VCs, founders & startup signals Daily signals on X & Substack Followed by Partners @ a16z, Khosla, Lightspeed, Bessemer!
Greg Turri @GregTurri
12 Followers 124 Following
kris @kollaid_
68 Followers 171 Following using ai agents to build ai agents | just a kid from chicago
Vaibhav Dixit @Vaibhavdixit02
330 Followers 2K Following
Dee Eff @flanny84
61 Followers 2K Following
z @_abdalazizz_a
283 Followers 562 Following
danialhasan @dhasandev
3K Followers 991 Following productizing agentic engineering @trysquadhq | manifest your destiny @_buildspace
Alejandro @Alejandro898837
116 Followers 5K Following Interesado en tecnologia, medio ambiente y me encanta leer. Aficionado a la programacion e Inteligencia Artificial. Boliviano🇧🇴
Ierofu @Ierofu9325
310 Followers 7K Following In order to be irreplaceable one must always be different. — Coco Chanel
Danial Namazifard @DanialNamazi_En
19 Followers 653 Following PhD Candidate in Computer Science at the University of Southern Denmark (SDU)
Marc @MarcMPSGC
102 Followers 1K Following
Dylan HadfieldMenell @dhadfieldmenell
5K Followers 3K Following Associate Prof @MITEECS working on value (mis)alignment in AI systems; Safety & Alignment Advisor at https://t.co/vt2gVrVr9f; @[email protected]; he/him
Anton @suganton
127 Followers 1K Following
CM @Creative_Math_
3K Followers 361 Following Deep Learning Research intern @blocks, grad student @UofT 🇨🇦 in a cool lab. Did pure math in a past life. cashmere-y
Pascal Thibeault @pascalthibeault
1K Followers 7K Following Tinkering with LLMs. Prev. lawyer @thefertpartners & @Medtronic
Evan @evan_posts
296 Followers 477 Following released teaser video for @withopine / https://t.co/IE58NMXVZk
Purav Gupta @puravg_
115 Followers 279 Following Bio-ML @tahoe_ai | @UofT @VectorInst | Dreaming about the Model of Life
joanna @joasyoo
1K Followers 668 Following applied ML @isomorphiclabs 🧬 prev. LLM scaling at @cohereAI 🇰🇷 🇨🇦 🇬🇧
Jaffar Keikei @jaffarkeikei
103 Followers 155 Following building @meconemarkets | prev google, uwaterloo, utoronto | backed by a16z speedrun making everything tradable
irfan sharif @irfansharifm
580 Followers 612 Following father figure. @cursor_ai, previously @modal, @cockroachdb
Josh Beyer @JoshBeyer8
65 Followers 348 Following I play music and code useful things https://t.co/ZFc1Df58Oy SWE with TD Securities and building Lapis
honnesh @create_compute
260 Followers 994 Following ai and human flourishing. writing poems mid-run. how can we deepen the human experience? eternally curious…
Fernanda Palacios @fernandaplcs
126 Followers 536 Following swe + ai + econ + finance. human life lover. prev. electric mind, twitter, uoft.
SabrinaEliot @d92Q6VDg949CR5A
63 Followers 1K Following
Matt Hardy @mdahardy
1K Followers 980 Following cto @proofofhuman, prev phd @princeton // language models, cogsci, ml
Aide @aidesupport
15 Followers 3 Following Deliver faster resolution, improved customer satisfaction, cost reduction, and intelligence that compounds.
Jim Banting @JimDBanting
1K Followers 2K Following Supporting partnerships, innovation, and entrepreneurship. Past: biotech bus dev (🇺🇸 & 🇨🇦) & co-founder. Views are my own.
Lengyue @LengyueMatrix
445 Followers 119 Following Founder of Fish Audio @ Hanabi AI Inc. Toward Personalized Voice Agent.
Fish Audio @FishAudio
25K Followers 19 Following AI voices you can't tell from humans. Expressive multilingual speech models and developer-first APIs. 👉 https://t.co/gzh4M4drSq @FishCreativeHQ
Speechmatics @Speechmatics
4K Followers 2K Following Smarter agents start with smarter listening. Build with the best ears in AI to understand every voice.
Daniel Lockyer @DanielLockyer
72K Followers 166 Following • I make sites faster and cloud/LLM bills lower 🚀 • 2:43 marathoner (🔜 2:39 in Berlin)
Ka @its6i77
2K Followers 558 Following lead engineer @HUMAIN | اللهم ارحم أبي | Majoring Cs at @uwaterloo
Jeff Dean @JeffDean
518K Followers 6K Following Co-founder & CEO of Discovery Loop. Former Chief Scientist, Google. Helped build many Google products, TPUs, Gemini, TensorFlow, MapReduce, Bigtable, ...
koray kavukcuoglu @koraykv
44K Followers 102 Following SVP, Google DeepMind Chief AI Architect, Google.
Discovery Loop @DiscoLoopAI
53K Followers 4 Following Automating discovery to accelerate science and engineering for the world.
Radical Ventures @radicalvcfund
7K Followers 235 Following Venture Capital fund focused on Artificial Intelligence and deep tech. Toronto, San Francisco, London and New York.
Oriol Vinyals @OriolVinyalsML
221K Followers 92 Following VP of Research & Deep Learning Lead, Google DeepMind. Gemini co-lead. Past: AlphaStar, AlphaFold, AlphaCode, WaveNet, seq2seq, distillation, TF.
Sanjay Ghemawat @Sanjay_Ghemawat
41K Followers 8 Following
Justin Uberti @juberti
18K Followers 130 Following Head of Realtime AI @OpenAI. Created WebRTC. Past: Distinguished Engineer @google, AIM. Amateur mathematician/musician.
Josh Whiton @joshwhiton
13K Followers 416 Following Solarpunk and Cybernetics. Building https://t.co/FixBRH6Ei4 & https://t.co/ZGZQknYq5P
Julian Salazar @JulianSlzr
2K Followers 649 Following Staff Research Scientist @GoogleDeepMind Frontier AI. Audio LMs (Gemini, Magenta RT2) & AI for math(.AG, .NT). Past: #NLProc #SpeechProc @AWSAI, @HarvardMath.
Andrew Ettinger @ahe23
1K Followers 2K Following CEO & Board Member @ Hume AI | 25 years building data infrastructure, now scaling emotional intelligence for voice AI
Rime @rimelabs
1K Followers 20 Following Rime voice models power the world's most important conversations.
Daryl W @Daryl_wanji
286 Followers 230 Following Engineer brain. Accidentally fell into mechanistic interpretability. Speak in weights & bias
Papers in the Park @PapersInThePark
8 Followers 6 Following Toronto ML Reading Group - no slides, minimal jargon, maximum curiosity - Sponsored by @Cohere_Labs
alvvin @alvanliii
453 Followers 3K Following this account is for shitposting, my other acc @alvanlii
Vals AI @ValsAI
17K Followers 270 Following Public LLM Evaluation // https://t.co/rNS39rhVQ9 @8vc @BloombergBeta @pearvc @a16z @HRTVentures @nextladder
Khurram Javed @kjaved_
6K Followers 194 Following Developing efficient algorithms for learning in big worlds @oaklab_ai Prev ~ Keen Technologies and PhD with Richard S. Sutton
Richard Sutton @RichardSSutton
70K Followers 59 Following Student of mind and nature, libertarian, chess player, cancer survivor. @ https://t.co/pioQVVjSXz, UAlberta, amii, https://t.co/vPRUv44glx, The Royal Society, Turing Award
Oak Lab @oaklab_ai
7K Followers 0 Following We are building learning algorithms that allow agents to achieve goals in big worlds.
Brian Johnson @code_brian
64 Followers 145 Following I’m thinking about creating Brian 2-5 to help me colonize space.
Mimi @mimihearing
2K Followers 1K Following Mimi is shaping the future of sound: intelligent, inclusive, and truly personal👂🎧
Partech @PartechPartners
30K Followers 3K Following A global tech investment firm headquartered in Paris with offices in Berlin, Dakar, Dubai, Nairobi, and San Francisco
David Zhao @davidzh
2K Followers 818 Following Co-Founder @livekit. Entrepreneur and engineer. I like computers and believe in hard money. #Bitcoin
Feiteng @FeitengLi
4K Followers 1K Following Speech · Image · Video · LLM multimodal generative models Research → Infrastructure → Built AI algo serving 250M+ users 公众号: Generative AI Open to opportunities
LiveKit @livekit
10K Followers 22 Following Open source framework and cloud platform for building voice, video, and physical AI agents. https://t.co/OWLvFH82oN
PolyAI @polyaivoice
9K Followers 186 Following PolyAI helps enterprises show up as the best versions of themselves in every conversation.
ai-coustics @ai_coustics
153 Followers 31 Following Building the future of audio intelligence. Real-time, AI-powered speech enhancement solutions. Ready to scale, for small teams and enterprise alike.
arXiv Sound @ArxivSound
7K Followers 38 Following Sound-related articles (https://t.co/dxVYgWJGOw and https://t.co/b90N0Zzvjs) on https://t.co/HHqPequzVU
Alexandre Défossez @honualx
5K Followers 541 Following Leading ambitious research @kyutai_labs. Chief Science Officer @gradiumai.
clem 🤗 @ClementDelangue
611K Followers 5K Following Co-founder & CEO @HuggingFace 🤗, the open and collaborative platform for AI builders
Hume AI @hume_ai
23K Followers 100 Following Build and measure voice models and agents the way people experience them, based on research and grounded in real human judgment.
Pipecat AI @pipecat_ai
5K Followers 4 Following 100% open source framework for realtime voice and multimodal AI. Maintained by @trydaily engineering team with support from the Pipecat developer community.
kyutai @kyutai_labs
27K Followers 14 Following
Gradium @GradiumAI
4K Followers 1 Following The voice layer for modern apps and agents. Real-time, scalable voice APIs: TTS, STT, turn-taking & voice cloning. Devs: build → https://t.co/r5CdNClhI5Trends for United States
You might like
