SaaStr’s data shows AI SDRs can send thousands of emails and boost responses up to 13‑times, but only when a proven playbook, ICP, and messaging already exist. Without that foundation, the AI merely scales zero results, ‑10x or 100x, the pipeline stays flat.
Tom Verrilli, chief product officer of Whatnot, argues that a lean team of senior product managers who stay hands‑on outperforms a traditional, ratio‑driven org. He explains how this philosophy, ‘we regret that product management exists’, fuels faster growth and better AI‑driven decisions at the live‑shopping platform.
The educator‑creator apologized for leaning on ChatGPT to script videos, warning that the dopamine surge from LLMs is "not healthy for me or good for the world." He’s cutting back on AI, pausing uploads, and urging creators to rethink dependence on generative tools.
Researchers unveil the 'Locksmith Loop', an agentic test‑synthesis technique that validates AI‑generated Java code against legacy COBOL by iteratively searching inputs and enforcing parity. On three case studies up to 4 k lines, it achieved near‑complete branch coverage and caught subtle migration bugs, offering a deterministic oracle for enterprise code rewrites.
Alibaba unveiled Qwen3.8‑Max, a 2.4 trillion‑parameter multimodal model with a 1 million‑token context window. Bloomberg reports it outperforms Moonshot’s Kimi K3 on several benchmarks, flirting with Anthropic’s Claude Fable 5. The launch pushes China’s AI race forward and gives developers a new high‑end alternative.
Johor’s government revoked Network School’s business licence on July 22, ordering the tech‑community in Forest City to shut down amid an immigration probe over alleged Israeli participants. Balaji Srinivasan responded by announcing a new MoU with Kazakhstan, positioning the venture for a rapid relocation.
You’re caught up.
That was all 6 stories in the 3 August issue.