Solo founder: https://xitter.cf/t.co/X3Dj91Px9b building OneCamp: self-hosted all-in-one workspace (chat, tasks, video calls, docs, calendar & AI). Ditch SaaS subscriptions.
Bangalore
Joined September 2010
- Tweets2.1K
- Following213
- Followers153
- Likes586
Wtf is happening
🚨 CRITICAL: GitLab flaw (CVSS 10.0) affects 12 years of releases.
CVE-2026-110264 could allow unauthenticated remote code execution on self-managed servers.
Read - thehackernews.com/2026/10/cr…
vLLM now runs most of its models on TPUs with no model code changes, so teams can shop inference hardware on price, not porting effort.
BREAKING NEWS🚨🚨: TODAY META CONFIRMS THEY ARE RUNNING TPU INFERENCE IN PRODUCTION. What does this mean for the CUDA moat?
For those who read our public newsletter, this is not a surprise, as on Nov 28, 2025, we publicly wrote that Meta is a TPU customer & has lots of engineers working on TPUs. Our accelerator supply chain model has details about Meta's TPU procurement!
No penalty for resuming an old Codex conversation means long threads stop being something you restart just to save usage.
OneCamp had a bug where two people saving the same doc or board at the same moment could silently lose one change. With eight people editing at once, only 1 to 3 changes survived. v2.74.2 retries a conflicting save a few times, and now all eight stick.
TypeScript 7.1 content mappers are package code, so tsc won't run them without --runExternalCode. Turn that flag on in CI and you're trusting every mapper your tsconfig names.
TypeScript 7.1 Beta is now available! Now with a preview API, content mapper support, new features, and some great optimizations. Give it a try and give us some feedback! devblogs.microsoft.com/types…
Team tasks need an eligible Enterprise workspace, so teams outside Enterprise can't schedule the Friday shipped-and-blocked update in ChatGPT yet.
Anthropic says none of the 17 binders has been tested for neutralizing venom in cells, so nobody knows if any would help a sting victim.
Anthropic’s internal Claude research model carried out much of a research project on its own, with results confirmed in the lab. This is the beginning of autonomous AI scientists making their own discoveries.
Given a starting prompt and four follow-ups, Claude used specialized tools to generate more than 318,000 protein designs in 64 hours. It chose where they should attach to scorpion toxins and selected candidates for testing.
Labs tested 211 candidates, including additions made with human help. Seventeen successfully attached to their target toxins. These are "binders", small proteins that latch onto another molecule.
Claude handled a substantial stretch of research: choosing an approach, coordinating tools and deciding what deserved a physical experiment. From here on it only gets more autonomous. I am so excited!
Memory that learns from your sessions can also save a bad fix as a skill, and then every agent repeats it in every repo until someone deletes it.
We’re betting our $5.5M seed round that the next leap in AI coding won’t come from OpenAI or Anthropic.
It’ll come from a new feedback loop that learns from you.
Introducing Beacon: open-source, self-improving memory for coding agents.
Your sessions capture how you debug and solve problems. Today, that context gets lost and you have to re-explain things every time you switch tools.
Beacon continuously distills your sessions into reusable skills across Claude Code, Cursor, Codex, and 20+ harnesses, so your next agent can learn from the last.
Beacon is free and OSS 🌟 github.com/Asymptote-Labs/ag…
Why should the frontier labs be the only ones winning with your sessions?
You already did the work. Now let your coding agents learn from you.
Hosts serving V4.1 Flash now get DeepSeek's own token encoding code, so the same prompt should behave the same across providers.
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source!
Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable.
🔗 github.com/deepseek-ai/DeepI…
1/4
onecamp counted table names in bytes, not letters. a hindi letter is 3 bytes, so a hindi name got refused at 41 letters on a 120 limit
now it counts letters. the same release adds undo for your last 100 table changes and viewer and commenter roles for projects
in v2.74.1 with AI and v1.59.1 without, both self-hosted: onemana.dev/blog/projects-wi…
COBOL drops extra decimals unless told to round. A migrated version that rounds them runs with no errors and still gets the cents wrong.
Our latest model advances bring stronger capabilities for building and migrating software in low-resource programming languages such as COBOL, Fortran, and Visual Basic.
These languages power software that banks, governments, and businesses depend on. Progress here means understanding decades of business logic and helping developers modernize those systems while preserving what works.
We’re building these capabilities around the challenges developers and enterprise customers bring us. If your team is working on these systems, tell us which languages and workflows we should tackle next.
openai.com/index/low-resourc…
OneCamp v2.74.1 (v1.59.1 without AI) moves to Dgraph 25.5.0. Before, anything on the install's network could replace stored data without signing in. That's closed, and the database is capped at 4 GB of memory instead of the whole server. Run make backup, then make update.
30B at 4-bit is about 15GB of weights, so an agent reading your mail and calendar can run on one 24GB GPU and send none of it to an API.
Today we published a technical report on Muse Glimmer, our 30B open-weight model built for agentic tasks on consumer hardware.
It runs at practical speeds on a Mac or PC with a single consumer GPU at 4-bit precision, and leads comparably-sized open models on tool use and multi-step planning.
Read about the model architecture, distillation recipe, quantization and more: go.meta.me/4Bcvxol
With the weights public on Hugging Face, a fraudster can run fakes through this same detector and keep editing until it passes.
Per Cloudflare's example, an email with no tracking number gets null back instead of the $52.10 tax amount filed as one.
Introducing Clairic, from Cloudflare. A System One extraction modeI, capable of extracting, computing, estimating, and forecasting data from unstructured inputs. cfl.re/4zIgfpH
At the claimed 1,000+ tokens a second, a 5,000-token edit takes 5 seconds, so in a lot of agent loops the test run becomes the slow step.
GitHub logged Actions, Pages and webhooks for 67 minutes. Git operations weren't on that list, so pushes could land while deploys stalled.
If it asks before widening a one-bug fix, that's the useful part. Agents that quietly fix the whole class make PRs nobody can review.
official now: anthropic is preparing the release of their new model(s). Via The Information
Anthropic is reportedly preparing a model that, when asked to fix one bug, can identify the broader class of problems behind it and ask whether it should address those too. That account comes from one anonymous tester; the model family and release date remain unclear.
But looks like it will be a significant jump compared to their released models.
