5 items5 builders

AI Builders Digest

What the people actually building AI said today. One page — a 2-min read.

Today’s sharpest idea was that reliable AI may come from small decisions embedded throughout software, rather than ever longer chatbot responses. Builders also weighed the promise of security agents against the human skills we could lose by leaning on them too much.

Podcast

The MAD Podcast with Matt Turck

JEV bets on reliable decisions inside software

Diogo Almeida argues that JEV’s fast, programmable classifiers can make small decisions that software can test and combine into larger workflows. A developer could, for example, set the confidence threshold for flagging an important email instead of trusting a chatbot’s polished reply. His concern with current models is reliability: “95% Einstein, 5% Mr. Bean. And if it doesn't tell you when it's the Mr. Bean, then how are you gonna automate anything?”

  • #models
  • #agents
  • #products
X

Aaron Levie

Box CEO

Aaron Levie sees a major opening for defensive cyber agents

Levie predicts vastly more tokens will flow through cyber agents as security teams struggle to keep up with expanding risks. He sees agents fanning out across tools to test attacks, inspect logs and assess enterprise data, creating opportunities across identity, networks, code scanning and data protection. He says recent gains in long running agents, computer use and cyber training make that newly plausible.

  • #agents
  • #security
X

Guillermo Rauch

Vercel CEO

Rauch says managing AI guidance is a core engineering skill

Guillermo Rauch says developers can overguide an AI that knows more than they do, or underguide one that lacks crucial context. Judging how much direction to give is the top skill he looks for. He also wonders whether the engineering language models use could be reduced to a much smaller working vocabulary.

  • #agents
  • #engineering
X

Ryo Lu

Ryo Lu worries agents could erase useful struggle

Ryo Lu says learning to build meant failing, asking others for help and developing judgment through repeated attempts. Agents can help people make things they otherwise could not, but a finished result can hide the questions they never learned to ask. He wants tools that help people share unfinished work and collaborate in the world beyond the screen.

  • #agents
  • #products
X

Peter Yang

Once Reading won Peter Yang’s family through a daily habit

Peter Yang says Once Reading is the only learning app he and his four year old have kept using for months. He credits its daily parent and child routine, a silly photo after each lesson, and physical books sent after 30 lessons. The books give his child a tangible reward for learning to read.

  • #products
  • #education

Get this in your inbox

One email a day. Unsubscribe in one click.

Where this comes from

Source data comes from the open-source project follow-builders by zarazhangrui, released under the MIT license. Summaries are generated by an LLM from that project's public feeds, and the summarization prompts are adapted from it. Every item above links to its original source.

Summaries generated automatically. Read the original before relying on any claim.