scaling @smallest_ai | Ex-@yugabyte DB | @ThePSF Fellow | community @bldrcollective try https://xitter.cf/t.co/uXFkpEsChH
LRU cache | BLR | SF
Joined December 2013
- Tweets14.5K
- Following5.5K
- Followers4.8K
- Likes78.5K
Pinned Tweet
docs are the first product a developer touches
so we rebuilt ours
six models, one place to build with them. go make something with voice
ps: increase the volume a bit :)
Abhishek (key/value) retweeted
Eminem's fastest verse vs Smallest Pulse STT
The last verse of "Godzilla" is one of the fastest in mainstream rap. Pulse kept up with all of it: 296 words, 1.4% word error rate.
Just the raw audio, transcribed live.
If Pulse can handle Eminem, your customer calls are easy.
Abhishek (key/value) retweeted
‼️ Researcher Erica Windisch skipped coordinated vulnerability disclosure and published 11 unpatched Tailscale flaws.
In her lab, Tailscale SSH race conditions let one hacked device get root on other servers and add a rogue node.
Readers added context they thought people might want to know
The Tailscale SSH findings require prior authorized access on the tailnet per the researcher's report; Tailscale stated issues concern post-authorization behavior only and did not recommend disabling SSH for everyone.
tailworm.com
x.com/IntCyberDigest…
Abhishek (key/value) retweeted
We have a background agent sweeping CI every hour, finding slow and flaky tests, optimizes the tests or code itself, and more importantly messages people on Slack if their changes introduced a CI speed regression.
The bottleneck has now moved back to the team in merging fixes the agent is producing.
The background agent framework is built on top of our internal software factory Foundry.
Abhishek (key/value) retweeted
Goodbye @haricm 😔
mediatinker.com/2026/10/21/h…
Abhishek (key/value) retweeted
marketing team needs a raise for this one lol
Abhishek (key/value) retweeted
technical founders trying to announce the birth of their child without captioning the post “hello world”
wtf, who made this smallest.ai
wtf, who made this techreferrals.in
Abhishek (key/value) retweeted
Editing takes skill. Captioning takes hours.
Capkit by Smallest AI takes the Captioning one off your plate. Click Generate and animated captions land on your Premiere Pro or After Effects timeline.
It runs on Pulse, the fastest speech to text model in the world, so the wait between Generate and captions on the timeline is barely a wait. 20+ languages, Hinglish included.
Link in the first reply.
Abhishek (key/value) retweeted
Catch our DevRel @StalwartCoder at BUILD Day Cochin on November 1 for a hands-on workshop on building your first voice agent.
Our 2nd speaker for BUILD Day is @StalwartCoder from @smallest_AI
He will be taking a workshop on AI: From Hype to BLDing your first Voice Agents Workshop. He is also passionate about Python, Voice AI & developer ecosystems. So, get ready for exciting tech conversations! (2/4)
Abhishek (key/value) retweeted
Our 2nd speaker for BUILD Day is @StalwartCoder from @smallest_AI
He will be taking a workshop on AI: From Hype to BLDing your first Voice Agents Workshop. He is also passionate about Python, Voice AI & developer ecosystems. So, get ready for exciting tech conversations! (2/4)
Abhishek (key/value) retweeted
Smallest 🤝 @hyperspell
Spent the afternoon with @contextconor at their office.
something's cooking 👀
also got some top tier merch in SF.
single till series B 💅
who’s at first?
🚀 Mistral AI’s Voxtral Mini Transcribe 3 ranks #2 out of 9 models by DER on Voice Arena’s Diarization + ASR leaderboard, outperforming its larger counterpart.
Following the launch of Mistral’s new Voxtral 3 family, we evaluated its Mini and Small transcription models on Voice Arena’s Diarization Bench.
🥈 Voxtral Mini Transcribe 3: 40.4% DER (#2)
Voxtral Small Transcribe 3: 41.1% DER (#4)
Both models were tested on 139 held-out conversations under identical conditions, using human-verified references, a strict 0 ms collar, and no predefined speaker counts or audio cleanup.
Mini leads on in-person conversations (40.6% vs. 41.5%), while Small performs slightly better on online calls (39.1% vs. 39.5%).
Overlapping speech remains the biggest challenge. Missed speech accounts for 82% of Mini's total DER. Both models correctly identify the exact number of speakers in 83.5% of sessions.
Congratulations to @MistralAI on the release ! 👏
View complete track & condition breakdown: voicearena.com/diarization-b…