@Hesamation I keep getting generic emails from them that they are working on adding additional capacity. But I don't know anyone who actually tested this.
English
Abhinav Chowdary
34 posts
Abhinav Chowdary
@BobbyCh03
Founder @RadiusAI | Building something exciting in healthcare | AI since 2017 | VoiceAI, Vision AI & agents | Proud father + husband | He/Him | Opinions my own
Katılım Nisan 2026
37 Takip Edilen20 Takipçiler
Remember this? 20 days ago SubQ claimed to have developed a model with 12M context window, 95% cheaper than Opus, and the same intelligence level.
they promised to release the paper and model card “next week”. that was 10 days ago. NOTHING.
the only update after this launch was a 3rd party eval by Appen, which mentioned it evaluated via Subquadratic API, and did not receive model weights, that’s normal, but not much of a proof.
if this is not a scam (which so far sets off every obvious red flags of being one) it’s super shady. you cannot make revolutionary research claims and still act like you’re running a sales pitch in a YC launch post.
the main problem is simple: they’re using a breakthrough claim to buy attention, credibility, and likely investor eyeballs before giving the community anything concrete to evaluate. very unprofessional compared to how an “ai lab” must conduct its research release.
Alexander Whedon@alex_whedon
Introducing SubQ - a major breakthrough in LLM intelligence. It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA), And the first frontier model with a 12 million token context window which is: - 52x faster than FlashAttention at 1MM tokens - Less than 5% the cost of Opus Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention). Only a small fraction actually matter. @subquadratic finds and focuses only on the ones that do. That's nearly 1,000x less compute and a new way for LLMs to scale.
English
@emanueledpt They have taken a thing or two from Oracles playbook. They want to lock people to their ecosystem. As AI model performance gap is tightening, model is no longer moat.
English
I’m so angry and disappointed in how Anthropic keeps fucking things up
I don’t understand why I can’t use my subscription with claude -p or Agent SDK
It truly doesn’t make any fucking sense
I even had a branch for Claude integration inside Remodex..
DPCode is becoming a better GUI/CLI wrapper everyday and now you have to go back to their shitty CLI
Luckily DPCode even supports multiple terminal tabs
And Remodex will support an integrated terminal too
But this is not the way to go forward
In the meanwhile
I’ll do my best to bring you a solution to these changes in my apps
ClaudeDevs@ClaudeDevs
Starting June 15, paid Claude plans can claim a dedicated monthly credit for programmatic usage. The credit covers usage of: - Claude Agent SDK - claude -p - Claude Code GitHub Actions - Third-party apps built on the Agent SDK
English
@plainionist I think its true atleast in our case. As a founder we always wonder if AI can replicate what we built. We ran bunch of tests trying to replicate our product it didn't even come close.
English
@antirez I 100% agree. Cant get anything done with opus 4.7. GPT 5.5 is so much better
English
I am the best OpenAI Codex Advertising, even if I would not sponsor a single video in my YouTube channel. But I believe on saying everybody when I think something is outstanding. I do it with restaurants, wines, persons I like, and AI systems.
antirez@antirez
Gentle reminder on how, in the recent DS4 fiesta, not just me but every other contributor found GPT 5.5 able to help immensely and Opus completely useless.
English
Abhinav Chowdary retweetledi
🚨 There's a major attack going on via npm right now.
Do not install any packages right now.
Talk to your agent ASAP and check if you're vulnerable or have been compromised.
This is urgent ‼️
International Cyber Digest@IntCyberDigest
‼️🚨 UPDATE: The TanStack npm attack is now a full campaign. 'Mini' Shai-Hulud has hit: - OpenSearch - Mistral AI - Guardrails AI -UiPath - Squawk packages across npm and PyPI The malware specifically targets AI developer tooling. It hooks into Claude Code (.claude/settings.json) and VS Code (.vscode/tasks.json) to re-execute on every tool event, long after the infected package is gone. npm uninstall does not fix this.
English
Just use Go/Rust/Zig/… anything but JS/TS
Socket@SocketSecurity
🚨 BREAKING: 84 TanStack npm packages were compromised in an ongoing Mini Shai-Hulud supply chain attack, adding suspected CI credential-stealing malware. Socket flagged every malicious version within six minutes of publication. This is a developing story.
English
@DanielSmidstrup Hello, would love to connect with fellow founders!
English
@theCTO Use it all the time especially for complex tasks, its an amazing harness
English
haven't met a single person that uses opencode
dax@thdxr
i've had a major shift in how much i use opencode in the past month it's tricky to articulate but seeing someone use it 10x more than made made something click what's weird is there's no fancy workflows or skills or anything, we're still just talking to the vanilla agent
English
@haider1 Totally agree, I have had better experience with codex since 5.4, 5.5 is even better
English
codex is so impressive right now
openai seems to be putting more focus on codex now, while claude models feel held back after the opus 4.6 update, especially because of compute limits
meanwhile, openai is resetting limits for fun
gpt-5.5 is genuinely the best model right now
English
@rezoundous maybe but I still see value in a co founder, it really depends on what you are building I guess
English
@justbyte_ they have different focus, they are not in coding per say, they are after search, research and task automation
English
@Kappaemme1926 really depends, I have Pro Plan and had to get another prod last night, but I do run things almost non stop. But i can say that Chatgpt Pro does offer you far more usage than Claude max, its not even close.
English
