Loveable Bags $200 Million in Funding, Now Valued at $1.8 Billion.......
Loveable, an AI company, secured $200 million in funding, raising its valuation to $1.8 billion.
The Cognitive Revolution·22 Aug 2025
Use this view to inspect the underlying evidence corpus. For ranked developments, decision posture and interpretation, use Signals.
Raw feed or Signals?
Raw feed is chronological evidence. Signals ranks and interprets material change.
Loveable, an AI company, secured $200 million in funding, raising its valuation to $1.8 billion.
Blue J uses GPT-4.1 with RAG to deliver cited tax research answers for US, Canada, and UK professionals.
NVIDIA released a 6 million example multi-lingual reasoning dataset for training and fine-tuning large language models across 30 languages.
SemiAnalysis compares NVIDIA H100 and GB200 systems, detailing power, TCO, and reliability for frontier model training, noting software improvements.
Hugging Face demonstrates integrating Claude 3 with image generation models, showcasing multimodal capabilities via API calls.
Hugging Face proposes a Multi-Modal Controller Protocol (MCP) to standardize AI interaction with research tools, enhancing agentic workflow integration.
Arm and ExecuTorch 0.7 aim to enable on-device generative AI, leveraging Arm's hardware for efficient model execution on edge devices.
OpenAI wrote to CA Gov. Newsom urging California to align state AI regulation with emerging federal and global standards.
Basis built AI agents on OpenAI o3, o3-Pro, GPT-4.1, and GPT-5 to automate accounting tasks, claiming 30% time savings for firms.
Standard Chartered has launched an AI-powered FX insights service for clients, claiming an industry-first in algorithmic market commentary.
Standard Chartered has appointed David Hardoon, former MAS Chief Data Officer, as Global Head of AI Enablement.
Standard Chartered has launched SC GPT, an enterprise-wide generative AI platform for its staff to assist with daily tasks.
Hugging Face introduced AI Sheets, a tool enabling data interaction and analysis using open-source AI models, similar to a smart spreadsheet.
The 'It Just Does Stuff' concept for GPT-5 suggests advanced autonomous agent capabilities, moving beyond task execution to independent problem-solving.
OpenAI releases GPT-5 via API with enhanced reasoning, developer controls, and improved coding benchmark performance.
OpenAI announces GPT-5 as its most advanced model, claiming enterprise AI, automation, and productivity improvements.
OpenAI publishes promotional content highlighting GPT-5 capabilities for coding and design use cases.
OpenAI released a demo showing GPT-5 performing workplace tasks including feedback analysis, product planning, and prototyping.
Hugging Face outlines new methods for aligning Vision Language Models (VLMs) using TRL, focusing on instruction fine-tuning and safety.
OpenAI published a case study on how code editor Cursor integrates GPT-5 into its AI-assisted development product.
OpenAI releases GPT-5 system card detailing a unified routing architecture across gpt-5-main, gpt-5-thinking, and nano variants.
OpenAI describes GPT-5's 'safe-completions' safety approach, replacing hard refusals with nuanced output-centric handling of dual-use prompts.
OpenAI publishes developer first-look video of GPT-5; no technical specs, benchmarks, or API details disclosed.
OpenAI launched GPT-5, claiming state-of-the-art performance across coding, math, writing, vision, and health tasks.
OpenAI released a video demonstration showing a ChatGPT agent conducting research, writing code, and taking web actions.
OpenAI & GSA offer ChatGPT Enterprise free to entire U.S. federal executive branch workforce for one year.
OpenAI has announced GPT OSS, a new family of open-source models hosted on Hugging Face for community and enterprise use.
OpenAI releases gpt-oss-120b and gpt-oss-20b as open-weight reasoning models under Apache 2.0 license.
OpenAI paper tests worst-case risks of open-weight GPT model via malicious fine-tuning in bio and cybersecurity domains.
OpenAI releases its most capable open-weights models, framing the move as a step toward broader AI accessibility.