GoKawiil
Tech News
An entrepreneur who runs a transcription company describes how handling sensitive audio and video files requires strict confidentiality practices, including cybersecurity tools, employee NDAs and clear client communication. The piece argues that trust, once built, can be destroyed instantly by a data breach, misleading claims or broken promises, and that honest acknowledgment of mistakes is key to recovery.
Otter.ai has secured $320 million in new funding and launched three AI-powered features aimed at extracting insights from meetings rather than just transcribing them. The move shifts the eight-year-old company's focus from real-time transcription toward workplace productivity tools.
Nvidia has launched Nemotron 3.5 Transcribe, an open-weight ASR model combining transcription, speaker diarization, translation and audio Q&A in one checkpoint, now available on Hugging Face and ModelScope. It transcribes 40 languages, translates 31 to or from English, identifies up to eight speakers across 13 languages, and handles recordings up to an hour long. Nvidia also released the Granary v2 training dataset and fine-tuning recipes via NeMo Speech tools to help developers adapt the model.
A new genomics study reports that variation in short tandem repeats—stretches of repeated DNA sequence—is associated with gene expression differences across distinct immune cell types. The findings, drawn from large-scale single-cell sequencing data, add tandem repeats to the known catalog of genetic variants that influence how genes are turned on or off in specific cell populations.
A transcription company founder writes that clients who pick vendors based solely on the lowest price frequently end up paying more when the work must be redone. The piece distinguishes price, cost and value, noting the author's firm regularly gets hired to fix botched transcriptions originally done cheaply elsewhere.
Google released AI Edge Foresight, a note-taking app currently available only on MacOS that transcribes meetings from Google Meet, Zoom or in-person conversations entirely on-device. It organizes notes, works with files like PDFs and Google Docs, and offers a Live Assistance feature to answer questions based on meeting transcripts, all without sending data to the cloud.
Soulog has released the Sense, an AI-powered voice recorder paired with the Soulog app that records, transcribes and organizes conversations by topic and decision. The app recognizes up to ten speakers, transcribes in over 120 languages, and can summarize conversations in a different language than they were held in. It also generates structured records of decisions and action items, plus a semantic map and timeline linking related discussions across time.
Mobvoi has released the TicNote Watch, a smartwatch focused on recording, transcribing and summarizing conversations without needing a paired smartphone. It uses dual omnidirectional microphones to capture speech from several meters away, offers up to 24 hours of continuous recording, and can store up to 200 hours of audio locally. The watch costs $249, though Mobvoi is offering it for $199 for a limited introductory period.
Researchers studied how somatic hypermutation, the process that diversifies antibody genes in B cells, is controlled by proteins that regulate RNA polymerase pausing and elongation, including NELF and the Integrator complex subunit INTS8. Using mouse models with Nelfb deleted specifically in germinal center B cells, alongside Ramos and CH12F3 B-cell lines, they mapped how these transcriptional regulators influence mutation patterns in immunoglobulin genes.
Google has rolled out a new setting in its Recorder app for Pixel devices called 'Allow recordings in other apps,' which is enabled by default. With it on, audio recordings now appear directly in a Recordings folder accessible through any file manager, rather than being sandboxed within the app.
MacWhisper released version 15.2 for Mac, adding live meeting transcription, improved speaker recognition using NVIDIA's Nemotron 3 Diarization model, and support for Alibaba's QWEN3-ASR model covering 22 languages. The update also deepens integration with Apple Intelligence, Siri, Spotlight, and Shortcuts on macOS 27, letting users search and open transcripts system-wide.
Apple's built-in Voice Memos app includes lesser-known features accessible by swiping up during a recording, including pause, rewind-to-replace, and multitrack recording on iPhone 16 Pro and newer. Users can also choose between Mono, Stereo, and Spatial recording modes in Settings, and access transcript copying, trimming, playback speed control, and noise-reduction tools from the recording screen.