logic tom
80 posts
is this the confirmation that whatever current dgx spark/ strix/ mac studio that exists in the near future will basically be able to run agi locally?
and hardware just becomes infinitely inflating?
PrismML@PrismML
Today, we’re announcing Bonsai 27B: the first 27B-class model to run on a phone. Bonsai 27B is the new multimodal flagship of the Bonsai family. Based on Qwen3.6 27B, it brings a new capability tier to local AI: multi-step reasoning, structured tool use, long-context workflows, and coherent agentic loops. Until now, models in this class have been impractical to deploy locally. A 27B model occupies roughly 54 GB in 16-bit precision, and even a strong 4-bit build is around 18GB - too large for a phone and for most laptops. Bonsai 27B changes that. It comes in two variants: • Ternary Bonsai 27B: 5.9 GB, 1.71 effective bits per weight, optimized for laptop-class quality. • 1-bit Bonsai 27B: 3.9 GB, 1.125 effective bits per weight, optimized for phone-class footprint. Everything is open-sourced today under the Apache 2.0 license.
English
@ashebytes obv very good video - but the music in the back is hitting.
English
bro is just beefing with everyone
Sam Altman@sama
come for the best model, stay because we don’t treat you with contempt
English
Will AI ads take over the NYC subway? Well, Michael El Toro is here to fight and he’s striking back hard🐂
English
anthropic by far has the best product on the market. they seriously need to rethink comms, everyone I know burnt through fable usage pending it going off sub.
only to get extended a week the day of.
this only incentivizes people to want to try the next like model
@ClaudeDevs
English
Which is why rn you should focus on solving problems/creating products where the way they sound/feel/look is totally beside the point.
Jared Friedman@snowmaker
LLMs are incredibly good at almost everything, except not sounding like an LLM.
English
homeboy you're the one sellling public market investors on short-term space datacenters
Elon Musk@elonmusk
He takes scamming to a whole new level
English
@Polymarket 😭 you're telling me i need to pay for perplexity / comet now
English
JUST IN: OpenAI to reportedly discontinue ChatGPT Atlas on August 9, less than 10 months after launching the standalone browser.
English
@sekachov @atlasformac Biggest ask right now is auto-tagging and correct labeling. Being able to separate typography from motion and graphical elements would be huge.
Additionally the ability to view, organize, move items across multiple folders is missing from a ton of tools.
English
@alogictom @atlasformac yeah, coming soon. what would you expect from it?
English
meet @atlasformac.
a native, local, and blazing-fast alternative to eagle app for collecting visual inspiration.
available now → asekachov.com/atlasformac
English
I’d also say the ability to focus on what they do best primarily.
Secondarily, Claude as a thought partner persona-wise is miles better. Grok behind it.
It would be naive to think they aren’t already internally using a Fable 5.2-5 or Mythos, and if labs are already struggling to 4.8 or F5 levels - I’d think moat is a lot wider.
English
This just might eat the market.
ClaudeDevs@ClaudeDevs
Claude Code on desktop now has an in-app browser. Claude can pull up docs, designs, or any other site. It can read, click through, and interact the same way it does with your local dev servers. It's sandboxed and configurable: you choose whether sessions persist.
English
i love this point in the cycle - where codex users rediscover a workflow that has been iterated in a million posts by Claude users for a month.
generally means anthropic is going to drop another model.
CJ Zafir@cjzafir
Look what I found guys. Here's how you can get 5x more Codex limts: > Use GPT 5.6 Sol (Extra High) as Orchestrator > Use GPT 5.6 Luna (Extra High) as Executors. Ask Codex to plan in depth, then spin up 3-5 GPT 5.6 Luna subagents and get more work done at 3x less cost.
English
Dumb question: why isn't there a Consumer Reports / Wirecutter for new AI model performance instead of relying on the labs selective benchmark model cards.
Standardized set of tasks / benchmarks.
Sort of like Simon Willison's SVG pelican test.
English