Neeren Chauhan, 'NC'

@astrange @simpsoka Enterprise software finally found its love language!

I strangely look forward to being greeted by my Codex pet each morning. The stickiest enterprise products might end up being the ones with the best consumer personality. h/t @simpsoka If you haven't built your pet - you're missing out on the fun and the added productivity: Keep working, get a friendly pet alert when your agents need your input.

Isomorphic out to prove AI can design drugs better than chemists. Is the pharma industry's "wait and see" phase is over? prnewswire.com/news-releases/…

VoltaGrid is selling time-to-power. If a data center can come online 2 to 4 years earlier by bringing its own gas-fired generation behind the meter, that speed may be worth more than the higher energy cost, emissions complexity, or infrastructure mess. @pmarca bloomberg.com/news/articles/…

India AI startup funding is down 21% YoY in 2026 so far tracxn.com/d/artificial-i…

Adobe launched an AI agent that lets you chat with PDFs and turn them into podcasts and social posts. solutionsreview.com/ai-news-for-th…

Claude is now inside Excel, Word, PowerPoint AND Outlook. It knows about your terrible formulas and your passive-aggressive email replies. There is nowhere to hide! anthropic.com/news/claude-mi…

AI vision agents use 45x more tokens than API agents for the same task. "Human-like browsing" turns out to be an insult when you're paying per token. reflex.dev/blog/computer-…

@mattpocockuk Efficiency = more intelligence, leverage, and action from every token you spend. You need outcome tagging at the call level. You need quality scores, not just completion rates. You need human-equivalent baselines. Most teams have none of these.

Way too many folks are concerned about total tokens When the right metric is value per token

@levie Token budgeting is the problem of this moment. By the time the tooling ships, the economics will have moved. You will be managing a cost line that is a rounding error.

A common trend emerging in larger enterprises is token budgeting as a major topic. As agents can do more and more long running tasks, and thus take vastly more compute, allocation of tokens across teams becomes a very real thing in the enterprise. Companies spend a meaningful amount of time deciding how much to spend on talent, marketing campaigns, events, laptop setups, and even the cost of lunches. Tokens will be no different. Tokens will similarly need to be excruciatingly well-managed because you’ll need to ensure you don’t blow up your budget, and you’ll need to ensure that the tokens are flowing to the highest and most useful parts of work. You don’t want to find out you burned your monthly budget on something relatively low value and then be blocked on the much higher value task later. Doing this at large company scale is extremely hard as you have layers of abstraction on data and visibility into the digital work being done by agents in any central way. This is going to mean that agentic spend will increasingly will expand beyond the confines of the IT budget, and end up in organizational budgets like other expenses. Ultimately team and org leaders will have to be given budgets for this, but even they don’t have adequate visibility and controls in most cases. We’ll need all new software just to solve this problem, and it’s probably an opportunity for startups in its own right. Going to be an all new era of enterprise resource allocation, especially while we compute constrained.

@joeschmidtiv In the current cycle, AI is taking the blame for decisions many companies already wanted to make. It is less the cause than the permission structure IMO. Most CEOs haven't even invested meaningfully in AI yet.

Critically important piece that was wonderfully done by DG Strong endorse!

@pmarca AI is taking the blame for decisions many companies already wanted to make. It is less the cause than the permission structure IMO.

Further, the employee activism movements of the last decade removed sentiment from many CEO's. The bill has come due for a tremendous amount of misbehavior.

@pmarca AI detectors measure pattern similarity, not authorship. Since AI was trained on human writing, a human can write something fully themselves and still get told: “Congrats, you sound like the robot we trained on humans.”

The true AI slop was the bad AI policymaking we met along the way.

@anothercohen not again!

Hantavirus preparedness checklist: ✔️ Toilet paper ✔️ Hand sanitizer ✔️ Toilet paper ✔️ Face masks ✔️ Sourdough starter ✔️ Ivermectin ✔️ PPP Loan

I think software is less like bread and more like media. Large studios did not disappear when YouTube arrived. They lost monopoly control over creation, distribution, and audience access. Then TikTok pushed it further: more personalized, lower friction, more creators, more formats, faster feedback loops. AI may do something similar to software. It will not make software “dead.” It will change who can create it, what gets built, how fast products ship, and where value moves. The harder question is whether we need the same number of people doing the same kind of software work, inside the same company structures, with the same economic model. The winners will be the people and companies who understand the user, the workflow, the edge cases, the trust layer, and the distribution. Simple products are still complex. AI just changes who gets to try building them.

Just because anyone can build software now doesn't mean software is dead. Anyone can bake bread in their home right now, yet 99% of us still choose to buy it from someone else. Simple products are complex! I will always be happy to pay someone to handle the nuances.

Private equity just became the AI distribution layer. techcrunch.com/2026/05/04/ant…

Mayo Clinic's REDMOD AI detects pancreatic cancer on routine CT scans up to 3 years before clinical diagnosis even when no tumor is visible. newsnetwork.mayoclinic.org/discussion/may…

Connecticut just passed AI regulation 131-17. Bipartisan. Not even close. Meanwhile the federal government is still debating definitions. At this rate we'll have 50 state AI frameworks before we have one national one. Welcome to the compliance patchwork. ctmirror.org/2026/05/01/art…

"The Deployment Company" sounds like the name of a villain organization in a mid-budget sci-fi movie!