agentic designer agentic coder
earth
Joined May 2025
- Tweets1.3K
- Following39
- Followers452
- Likes75.4K
Apparently, the usage limits for Grok Bot were decimated without warning or even an announcement, with Elon blaming the costs relative to revenue 😬
It's so sad seeing how big of an uplift in performance more VRAM can give particularly in higher res gaming
yet, the amount of VRAM graphics cards ship with rarely budges from one generation to the next, and the VRAM on a graphics card is never officially user upgradable... 🥲
Modded RTX 2080 Ti with 22 GB memory delivers up to 73% boost at 4K compared to 11 GB variant, benchmarks reveal. 🔗 wccf.tech/1lh0u
Not only is OpenAI bringing Codex's usage limits to ChatGPT, but they're also limiting how many paid accounts you can use on 1 device, forcing users to pay for credits (which give far less usage per dollar than a sub) instead for more usage
According to Redditors, OpenAI has now sent an email about ChatGPT & Work merging, + limits sharing between ChatGPT and Codex as a result
It appears only Instant will be unlimited now
the whole email: xitter.gq/liuzhihangs/status/211…
OpenAI sent an update email. Let's break down the real message:
1. "We won't reduce your Pro 200 limits." → Thanks for not nerfing what we already paid for. (But Pro 100 still doubled, so our relative advantage is gone).
2. "Chat and Codex will share the same allowance." → So using Codex will drain my Chat quota? A huge downgrade in disguise.
3. "If you run out, buy credits." → The ultimate goal. 🙄
They took away our 4x advantage, merged our quotas, and now want to sell us credits.
it appears there's a GPT-6 Rosalind Research model now for those with Rosalind Research access
🚨 GPT-6 ROSALIND IS REAL!
OpenAI has quietly released a specialized GPT-6 model for biological reasoning and complex life sciences research.
API pricing:
$15/M input tokens
$75/M output tokens
Access is restricted to approved research organizations.
This isn't a leak anymore. It's officially in OpenAI's API changelog!
Damn! Anthropic renamed the Max plans, but more importantly, Ultra (previously Max 20x) has no 5-hour limit now! 👀
We’re renaming our Max plans:
Max 5x → Max
Max 20x → Ultra
Max is still $100/month, and Ultra is still $200. If you’re a subscriber, there’s nothing you need to do.
We’re also removing the 5-hour limit for verified Ultra accounts, starting today with most existing subscribers. Newer accounts keep a temporary limit until they’re verified, which helps us prevent fraud and abuse. You'll get an email when it's removed.
Weekly limits remain unchanged. View our plans here: claude.com/pricing
If you have a Claude subscription, you can now use it in OpenCode again (they got permission from Anthropic for it this time!) 🥳
🚨 BREAKING: ChatGPT potentially to share usage limits with Codex as soon as December (could always change of course, not yet officially announced)
personally, this gets rid of the only advantage ChatGPT plans still had to me over Claude subs (I don't use image gen etc) 🥲
OpenAI is no longer going to slice ChatGPT Pro 200's usage limits in half on the 30th, instead opting to keep its existing 20x usage limits, while *also doubling* the usage limits for Pro 100 & Pro 500 👀
Day 25/
We are doubling all usage limits for all Pro 100/200/500 tiers and Business Premium, effective today.
We heard your kind feedback and won’t be reducing usage limits for Pro 200, so your Work and Codex usage will stay the same. We’re also going to continue offering extended limits for dots for now, and usage of your dot will still continue to be separate from your default usage.
And we'll of course also keep shipping better models and other efficiency improvements that increase your effective usage over time.
ah yes, a zero-click denial of service vulnerability that can permanently (persisting even beyond factory reset!) disable phone calling on an Android phone isn't a security issue according to Google 🤦
‼️ Google won't fix an Android bug that lets one invisible text crash a phone's calling service, researcher Mark Blaszczyk says.
So he just released the Proof-of-Concept!
The malformed message never reaches the inbox and needs no user interaction. It drops active calls, emergency calls included.
He says about 30 sent in a row corrupt the phone's SMS database, so a Pixel 9 or Samsung Galaxy crashes on every boot. A factory reset only helps until the carrier delivers the texts again.
He reported it to Google on 5 May. Google closed it as "Won't Fix" and a "Non-Security Issue".
Readers added context they thought people might want to know
Google will fix it as part of the November security update. A fix has already been pushed to AOSP in October, with the first commit in June.
x.com/GrapheneOS/sta…
I think the move to simple design languages with 0 detail whatsoever just as high PPI displays were beginning to hit the market was a terrible tragedy 😔
algorithon retweeted
The computer science collaboration will understand and explain a result concerning the Exponential Time Hypothesis (ETH), a longstanding conjecture about the computational complexity of Boolean formula satisfiability, a problem deeply related to the P vs. NP problem.
Surprisingly, our system has recently proved that the ETH is false, a result that was subsequently verified in Lean.
The refutation of the ETH reopens questions about how much faster these tasks could become, and may help guide the search for better algorithms. Finding such algorithms, and turning theoretical advances into practical improvements, will require significant further research.
Our system found its natural language disproof for the ETH we reached out to to @rrwilliams. Ryan will form a collaboration with Russell Impagliazzo and Ramamohan Paturi to verify and shepherd this result. Impagliazzo and Paturi, along with Francis Zane, are the original proposers of the Exponential Time Hypothesis, and Williams has refuted strong versions of the ETH, used ETH for hardness results, and connected algorithms and lower bounds.
This group asked us to make all raw data of the results public to facilitate independent verification by the community, which we have done by posting the natural language proof and Lean formalization to our GitHub repo. The group also invites the community to help organize an open examination of the ETH claim and to discuss the broader questions that AI-generated breakthroughs raise about verification, access, credit, and the future of research.
github.com/openai/math/blob/…
I've been obsessed with UI design for many years
when I was a child, I actually wanted to be a UI designer someday!
I can come up with *ideas* in my imagination that I like, but I struggle to translate those into real designs in Figma & other tools
not with Claude Design! 😁
hmmm... I wonder if this implies Anthropic is expecting the Claude models based on their next base model to be *significantly* more capable than the 5.5 series? 🤔
After switching to Claude again 'cause of Opus 5.5, I was disappointed that Claude Chats share a usage quota with Claude Code, unlike ChatGPT where Chat is separate from Codex/ChatGPT Work,
after learning how well Haiku 5.5 works for my chat use cases, I'm no longer disappointed
literally every idea I come up with is being executed on already by someone else who had the same idea but with more money to spend on tokens than me 😭
