DreW

Lemonade ➡️

[DreW @devteamdrew]

Lemonade 2.0 is officially live.

https://t.co/Ehb9oWdpJi new official website has launched.

Studio is here.

An infinite free-form canvas for your AI agents.

Drag terminals, sticky notes, browser nodes and media nodes anywhere.

Run up to 16 agents simultaneously.

Claude Code, Gemini CLI, and Codex on one screen.

$29 one-time payment.

No subscriptions.

Every future update included forever.

Stop switching tabs and start shipping. 🍋

[DreW @devteamdrew]

@ClaudeDevs It worked! Thanks @bcherny

https://t.co/FFbamwZCIh

[DreW @devteamdrew]

POS workflow feature from @claudeai just used my entire 5 hour usage on a research task that I didn't ask for directly - just wonderful

Running 104 agents to answer a simple DNS question

@bcherny @trq212

[DreW @devteamdrew]

Just ask AI to build it 😎

[b wall @bwallxyz]

just do it

[DreW @devteamdrew]

@Aditya_181105 Sonnet 4.6

[Froots @frootsai]

Froots:
- Infinite agents
- Infinite databases
- Agent browser
- skill library
- routines that connect to widgets
- internal kb
- internal memory
- no additional downloads on install

Hermes agent desktop:
- cli wrapper
- browser and memory are external features that download on install
- memory is external
- no built in Knowledgebase

[DreW @devteamdrew]

@frootsai this looks insane dude.

UI and UX is perfect🤞

[DreW @devteamdrew]

@bridgemindai only thing is it'll take more CPU and RAM resource to run since its now a GUI rather than a pure CLI experience, I find with Hermes CLI I can run 4/5 of them on an 8GB RAM Dell Optiplex in-tandem

Definitely a use for both GUI and CLI experience

[DreW @devteamdrew]

@nishithX26 I use https://t.co/EaSvbkvEkk now, it’s built for agentic development solely. No bloat

[DreW @devteamdrew]

@coderaw_ Claude > Codex overall

[DreW @devteamdrew]

@TTrimoreau Real frontend design, these models all give the same half-arsed UI attempts

[DreW @devteamdrew]

@suni_code https://t.co/EaSvbkvEkk

[DreW @devteamdrew]

@bridgemindai Made my own GUI for Hermes and I’m loving it

[DreW @devteamdrew]

POS workflow feature from @claudeai just used my entire 5 hour usage on a research task that I didn't ask for directly - just wonderful

Running 104 agents to answer a simple DNS question

@bcherny @trq212

[DreW @devteamdrew]

@KaiXCreator It’s quite simple really, those who teach themselves new skills will have a far better time with AI - better prompting, better understanding, better design skills

AI sucks at frontend design

[DreW @devteamdrew]

@JustJerry121 Thank you! It’s designed for all levels of builder

[DreW @devteamdrew]

Builders!

I'm trying to get to 100 followers today

I want real builder people, not random numbers

If you are building with AI agents, devtools, boring SaaS, SEO, ecommerce ops, or weird practical stuff

Drop your link below!

[DreW @devteamdrew]

@trq212 @sidbid @trq212 would you say we're close to a complete solution for agentic engineering?

[DreW @devteamdrew]

SOTA is 1.5%

Crazy.

[ARC Prize @arcprize]

Anthropic Opus 4.8 is new SOTA on ARC-AGI-3

Score: 1.5%, ~$10K

ARC-AGI-3 analysis notes:
* Opus 4.8 read the environment an abstraction *above* Opus 4.7, as objects & systems, not pictures
* Opus 4.8 succeeded on early levels, but still committed to a wrong sub-goal

[Massimo @Rainmaker1973]

Multitasking vs focus

[DreW @devteamdrew]

@FlippedRay Gemini is a general purpose model it’s only losing the coding race (probably the most important right now though)

[DreW @devteamdrew]

@TTrimoreau

[DreW @devteamdrew]

I think @Google’s biggest mistake is using Gemini to build their products

Using a general purpose model for high level coding tasks is no bueno