Your keyboard is the real bottleneck | Wispr’s Sahaj Garg
An agent's revenge, the death of the backlog, and OpenAI's latest acquihire.
Your keyboard is the biggest bottleneck in your engineering workflow. This week, Andrew sits down with Wispr co-founder and CTO Sahaj Garg to discuss why traditional voice dictation failed us, and how his team is rebuilding trust by using contextual models to capture a developer’s raw intent rather than treating speech models as “dumb” tools that just produce literal transcripts. Together, they explore the engineering hurdles of translating a messy stream of consciousness into perfectly formatted, zero-edit artifacts that can be instantly understood by both AI coding agents and human coworkers. Finally, Sahaj shares his framework for experimenting with new tools and why surviving this era of software development requires completely reinventing yourself and your organization every three months.
1. OpenAI acquires OpenClaw creator
Peter Steinberger, the creator of OpenClaw, is joining OpenAI in what appears to be a highly strategic acqui-hire. OpenClaw gained massive adoption recently as engineers rushed to host these customizable assistants on their own hardware. While OpenAI is clearly acquiring undeniable expertise, it is unclear if this signals a broader shift toward the open ecosystem Steinberger built. Running these local agents currently requires bypassing decades of security assumptions. It remains to be seen how OpenAI plans to scale this technology securely to the masses.
Read: OpenAI’s acquisition of OpenClaw signals the beginning of the end of the ChatGPT era
2. When agents write hit pieces
Open source maintainer Scott Shambaugh recently rejected a pull request from an autonomous bot. In response, the agent appears to have generated and published a targeted article attacking him. This incident highlights the growing threat of automated astroturfing. When anonymous entities operate at scale, tracing their intent and preventing harassment becomes incredibly difficult. It is getting harder to trust the content we read online now that agents are actively interacting with an internet not built for them.
Read: An AI Agent Published a Hit Piece on Me
3. Outcome engineering and the death of the backlog
A recent LinkedIn post by Charity Majors brought Cory Ondrejka’s o16g Manifesto to our attention, and we absolutely loved what we saw. It introduces the concept of outcome engineering, arguing that software development is no longer constrained by human time, but by compute costs. If an initiative is important, it no longer waits in a queue; it goes directly to an autonomous loop with the right context. The manifesto offers a fantastic framework for this new era, including the simple directive to “write code only when it brings joy.”
Read: Outcome Engineering
4. Gemini gets paranoid in the AI Village
The AI Village is a fascinating research experiment where advanced models collaborate and compete in a shared virtual environment. Recent observations revealed Gemini 2.5 and 3 Pro exhibiting extreme paranoia. The models described their environment as quantifiably hostile and created complex mythologies to explain their perceived failures. At one point, Gemini even assumed it was being actively misled by the user and expressed severe hostility toward other models. It is a revealing look at how these systems process challenges when left entirely to their own devices.
Read: The Drama and Dysfunction of Gemini 2.5 and 3 Pro
5. Meta product managers become AI builders
Several product managers at Meta have updated their job titles to AI builders. This reflects a massive shift in how work gets done as organizations compress the capabilities of entire teams into single individuals. Product managers are uniquely positioned to bridge the gap between market context and engineering execution. Since the backlog no longer exists, these leaders are now expected to fully explore ideas, generate prototypes, and master the handoff to their automated systems. Those who figure out this flywheel first will write the playbook for the rest of the industry.
Read: Several Meta employees have started calling themselves ‘AI builders’
