Sean Kerner
@TechJournalist
Followers
66K
Following
125K
Media
7K
Statuses
49K
IT consultant, technology user, tinkerer and sometimes Klingon ; tips [email protected]
online
Joined July 2008
Sean Michael Kerner's Spring 2023 Human Writer certificate https://t.co/zBLW0YnjcW That's write I'm human. I write about #AI. Skynet hasn't won (yet).
authory.com
This certificate proves that the holder's articles are written with a high degree of certainty by a human, not by artificial intelligence.
6
4
32
NEW — remember Scale AI? Six months after Meta poached its CEO: • Valuation has fallen to $7.3B one one platform, down from $29B in the Meta deal. • Pay and workloads for AI training tasks have been slashed. • Activity in the main chatroom for AI trainers has plunged.
1
1
1
How Hud's runtime sensor cut triage time from 3 hours to 10 minutes https://t.co/MsCWdJMe4I via @VentureBeat "Every software team building at scale faces the same fundamental challenge: building high-quality products that work well in the real world," @roee Adler, CEO and
venturebeat.com
0
2
3
Enterprises finally get a heads-up on what AI-generated code actually does in production https://t.co/VRT1xKBi2H
venturebeat.com
0
4
15
The problem with small models is that they are very narrow. Even if they are not built to pass a benchmark, they lack the general reasoning, resilience & knowledge of larger models. That may not matter for some tasks types, but you don't want a customer service AI to be naive.
10
8
148
Databricks builds enterprise-focused OfficeQA AI benchmark after finding academic tests miss real-world document tasks https://t.co/qX7rZUpMuC
venturebeat.com
0
4
14
Introducing AlphaEvolve, our Gemini-powered coding agent for designing advanced algorithms—in private preview. AlphaEvolve uses LLMs to propose intelligent code modifications, creating a feedback loop that evolves your algorithms to be more efficient → https://t.co/PQtYUtOtkN
58
391
3K
🚀 We introduce Soft Adaptive Policy Optimization (SAPO) — a smooth, stable, and highly effective RL method for training large language models. Why SAPO? 🔹 Hard clipping is brittle — gradients vanish or explode 🔹 MoE models amplify variance, making training even more unstable
arxiv.org
Reinforcement learning (RL) plays an increasingly important role in enhancing the reasoning capabilities of large language models (LLMs), yet stable and performant policy optimization remains...
26
165
1K
✉️Shipping an email agent now just requires a prompt. Using LangSmith Agent Builder, you can create an agent to: - Prioritize and label inbound messages - Draft responses - Run on a schedule or on demand Email agents are already one of the most popular Agent Builder use cases.
6
19
112
This seems right today. I think what will change is when models start inking deals for exclusive content. Just like YouTube started as a technology play and became a content play, so too will LLMs.
LLMs are the new disk drives: commodity infrastructure you hot-swap for whoever’s cheapest + best. The fantasy that the model is a moat just expired. ☠️
15
8
108
Some notes on the two features in the newly released Django 6 that most caught my eye: background workers and template partials
simonwillison.net
Django 6.0 includes a flurry of neat features, but the two that most caught my eye are background workers and template partials. Background workers started out as DEP (Django Enhancement …
5
5
99
Ditch the silos. Build on the platform that truly scales. 🚀 The Snowflake Startup Challenge is on! We're scouting the best new apps powered by the Snowflake AI Data Cloud. Up to $1M in funding + a 1:1 session with CEO @RamaswmySridhar is on the line. Apply now 👉🏻
2
2
8
Safety-focused minimal distros like Alpine and Wolfi have new competition. @RedHat introduces Hummingbird, its take on streamlined, enterprise-ready container images. By @sjvn
https://t.co/Ajrnj9evVA
thenewstack.io
Safety-first, minimal Linux distributions, such as Alpine Linux and Wolfi, have become popular, and now Red Hat is entering the field with Project Hummingbird.
0
2
2
How are the best teams actually building with AI coding agents? Join Harrison Chase and Jason Ginsberg (Head of Engineering, Product at Cursor) for a fireside chat on: - Coding agent UX - How teams are building with Cursor - Trends heading into 2026 This will be an in-person
2
13
68
venturebeat.com
1
2
2
26,000 users in a blind test picked @Google Gemini 3. The case for evaluating AI on real-world trust, not academic benchmarks "It's the consistency across a very wide range of different use cases, and a personality and a style that appeals across a wide range of different user
venturebeat.com
1
1
3
This is the momentum check. Agentic AI just got real: • @AnthropicAI models live in Snowflake Cortex AI. • $2B+ in @awscloud Marketplace sales • @Accenture just launched a Snowflake accelerator • 1,200+ customers already deploying Snowflake Intelligence We're where
Big day at @Snowflake! We’re accelerating enterprise AI adoption and building serious momentum. Here’s a quick update on some of our latest news. 👇
2
6
30
Gemini 3 Pro scores 69% trust in blinded testing of 26,000 users: The case for evaluating AI on real-world trust, not academic benchmarks https://t.co/KGJjzH5N7s
venturebeat.com
1
3
9
Everyone uses Amazon Web Services (AWS) S3 for cloud object storage... But will everyone use it for vectors too, now that Amazon S3 Vectors is GA? What does it mean for vector database vendors? ->with insights from @awscloud @holgermu @Gartner_inc and more...
venturebeat.com
0
0
3
AWS claims 90% vector cost savings with S3 Vectors GA, calls it 'complementary' - analysts split on what that means for vector databases https://t.co/khsrAae2vo
venturebeat.com
0
2
8