super cool! i just started looking into the intersection of computer vision and healthcare. would be interested if you have more details on how you built it and what went into it or if you even have a repository with the code. i'd love to look into that. either way super cool project!
Introducing SID-2, a new search model built on compute-scaling swarms.
Our swarms find better results than Opus and Astra, while being 8x faster and 20x lower cost.
By 2030, AI will write an internet's worth of text every 24 hours. Current search pipelines were built on hand-designed rules, adding more won't scale to meet this incoming volume of data.
SID-2 is available on our API today. Try it on your most demanding searches.
probably closer to reality: @joelniklaus re-ran Hermes offline, so its 166 DeepSWE passes from online lookups no longer count. It drops from 71.0% to 61.9% on DeepSeek-V4.1-Flash.
DSH is now 2nd at 70.6%, just behind OpenCode 2 (70.8%), and 1st at high effort (71.7%).
Love the idea of this but I think Hermes’s lead isn’t real here, once its lookups are corrected DSH looks better. It solves terminal tasks better, is more consistent and is faster, at a slightly higher cost.
https://t.co/DUOYtJJox8
Would it be fair to characterize the state of enterprise AI adoption as:
"We don't really know if we're getting any return from all this spending, but our employees would lose their shit if we took the tools away?"
This is hands down the best tech purchase I’ve made in the last year.
Charges phones and laptops very fast and many times over
Fits in a jacket
No extra cables to remember
$120.
Asked claude to rebuild @MaMoMVPY’s chart from OECD, SNB, IMF and BIS data. same result: 46 economies, slope 0.52, R² 0.59.
But switzerland backs your QE point. The snb balance sheet hit CHF 1.06trn and base money is now below its 2019 level. M3 grew 2.2% a year. Inflation: 1.1%.
Docs say neither includes code, prompts or file paths. Turn both off in ~/.claude/settings.json under env:
"DISABLE_TELEMETRY": "1",
"DISABLE_ERROR_REPORTING": "1"
Or CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1 for all of it.
Just noticed that Claude phones hole A LOT: My Pi-hole blocked 2,421 attempts in a week alone.
It sends usage metrics and error reports to Datadog by default:
• https://t.co/cElZzopj3i
• https://t.co/DXNhkzV8c5
Meet Ring Drone Cam 🚀
Unlock a new way to take command of your home. Fly room to room for unobstructed aerial views. Use Live View to check in on different rooms or create custom patrol routes unique to your space, right from the Ring app.
Learn more: https://t.co/4jqxa7NevQ
Remember when anthropic killed jp morgan, goldman sachs et al by releasing ten agent templates “for the most time-consuming work in financial services”
I have ~25-30 agent sessions, running on different machines, working on different projects, using different harnesses ...
all managed natively from macOS using Rex
I think Hermes’s lead is inflated: its score still counts 166 DeepSWE passes where it looked the answer up online. Corrected, I would estimate it around 65.5%. DSH would then sit well above it. @joelniklaus Do you intend to fix that at some point?
Love the idea of this but I think Hermes’s lead isn’t real here, once its lookups are corrected DSH looks better. It solves terminal tasks better, is more consistent and is faster, at a slightly higher cost.
https://t.co/DUOYtJJox8
DeepSeek V4.1 Flash 1017, GLM 5.3 Flash, Qwen 3.8 Next Flash, and even Qwen 3.8 27B are all outperforming (in both intelligence and capabilities) every model that was considered "frontier intelligence" like 11 months ago.
I think open source ai is going to be fine
another brutal reality check for the open-sauce china shills
GPT-5.5 released on April 23 outscores all tested chinese models
and OpenAI and Anthropic have been holding back their strongest models since the supply chain risk designation, so we aren't even seeing the full extent of the domination
Love the idea of this but I think Hermes’s lead isn’t real here, once its lookups are corrected DSH looks better. It solves terminal tasks better, is more consistent and is faster, at a slightly higher cost.
https://t.co/DUOYtJJox8