maelus

can’t compare OpenAI / Minimax / etc to Anthropic, they want their customers, 100% different prisms they wish they had as many users as Anthropic. They want new users. meanwhile Anthropic struggles to meet the demand as they’re growing a LOT - and want to favorise their native services users. don’t you think it makes sense? Once their infra is scaled maybe they’ll change again, who knows

@abustamante @davemorin Funny how it works for literally any other player in the AI industry, we support subscriptions from MiniMax, Alibaba, OpenCode, GLM, OpenAI...

woke up and my mentions are full of these Both me and @davemorin tried to talk sense into Anthropic, best we managed was delaying this for a week. Funny how timings match up, first they copy some popular features into their closed harness, then they lock out open source.

@mauricekleine what’s your current conversion? Have you tried hard paywall with like 1 week trial? (It’s what boosted my conversion one of product - for your product that could makes sense / fit as there are no AI cost involved etc so basically it costs you nothing to give one week for trial)

here’s how we increased new Mockly signups by 350% in 4 weeks: initially, the signup screen made users sign up AND pick a plan in a single screen many dropped off now, users sign up first and pick a plan on the next screen they might still drop off then, but now we can start an email campaign to try to get them to convert in the days after signup small tweak, big results ps yes aprils fools day saw a peak in signups, but pranksters don’t convert! took us 9 months to get to 2000 users, and then we added 1297 in the course of 4 weeks upwards and onwards 📈

Le “seul” élément qu’il manque pour que ce soit l’avènement des robots - c’est le stockage d’énergie. Tant qu’on a pas mieux que les batteries lithium ions, ça va galérer - du coup imo la vrai rupture arrivera quand il y aura un rupture sur les batteries (qui décalera la prochaine rupture nécessaire à la génération de l’énergie haha) Well done pour ta video, très stylé de te voir expérimenter sur le délire humanoid!

Svp regardez ne serait ce que le début de ma dernière vidéo YouTube. On se rend pas compte du truc en marche actuellement. C’est le moment de monter dans le train.

@nikitabier Love everything you’ve been doing - this one feels restrictive, non-us users with international audience will get lost I’m French, I never post in French and don’t wanna focus on French audience only

Starting Thursday, we'll be updating our revenue sharing incentives to better reward the content we want on X: We will be giving more weight to impressions from your home region—to encourage content that resonates with people in your country, in neighboring countries and people who speak your language. While we appreciate everyone's opinion on American politics, we hope this will disincentivize gaming the attention of US or Japanese accounts and instead, drive diverse conversations on the platform. We invite creators to start building an audience locally. X will be a much richer community when there's relevant posts for people in all parts of the world.

@trq212 how do you handle multiple choice for AskUserQuestion and PlanMode etc?

We just released Claude Code channels, which allows you to control your Claude Code session through select MCPs, starting with Telegram and Discord. Use this to message Claude Code directly from your phone.

@jackfriks @postbridge_ MCP was genius move

okay grew 40% last month, now ARR is up to over $250,000 USD so that mean @postbridge_ is worth $1,000,000 :D

@joemasilotti Expo but for RoR

What if Hotwire Native apps didn't require you to open Xcode or Android Studio… at all? What if a Rails developer could write 100% Ruby, HTML, CSS, and JavaScript to build their mobile apps? What if they could submit these apps to the app stores… automatically? 🤔

we’ll see a shift in coding test won’t make sense anymore good context + ai agent coding + ai automated issue tracking & solving >> 200k lines of test imo

@steipete @openclaw weird to to see GLM 5 < GLM4.5 air and Minimax M2.5 < Minimax M2.1 that feels off

Interesting benchmark on which model is best for @openclaw pinchbench.com

@levelsio force using opus 4.5, less dumb mistakes somehow /model claude-opus-4-5-20251101 you can add it in the settings.json so it always start with it

Claude Code with Opus 4.6 was so dumb today I finally had to write my own code again A sad state of affairs 🥹

@trq212 @LinusEkenstam what model are you using? Is it same as for Claude app? found it not as much accurate as Whisper v3 / ChatGPT tts for french in the past (haven’t tried much since as it was often getting things wrong)

Voice mode is rolling out now in Claude Code. It’s live for ~5% of users today, and will be ramping through the coming weeks. You'll see a note on the welcome screen once you have access. /voice to toggle it on!

@esprit_dev full opus, et opus 4.5 plutôt que 4.6 pour moi, moins d’erreurs bêtes

Avec Claude code vous êtes plutôt full Opus ou vous alternez Sonnet et Opus pour certaines tâches ? Bon j'imagine qu'avec l'abonnement Max, tu peux utiliser Opus à fond. (J'ai le pro ).

"use a council of agents to analyse my request, your plan, and give you feedback. First, self-reflect about what you want feedback / additional inputs / approach about, think in steps for that (UX? data? implementation logic? code approach? etc) - summon two agents with the exact same prompts/instructions: Opus 4.6 and Sonnet 4.6. Then analyse their feedback and self-reflect. Ultrathink"

claude code hack: 'council of agents' during plan mode i'm using Opus 4.5 as default, and sometimes using a `council of agents` eg telling Opus 4.5 to summon Sonnet 4.6 & Opus 4.6 and ask their take on a subject -> improves/verify the plan for critical subjects prompt below

can someone explain why @Apple is not rolling out Markdown *writing* for Apple Notes ??? fr, that's the ONE thing missing

using termius: “Review and upload files with SFTP To review files generated by your agent, open an SFTP tab directly from your terminal without leaving your session. You can use the same SFTP tab to upload screenshots of a website, then include those files as context in your agent's prompt. Your AI agent can analyze the uploaded files and generate the required updates or a similar design.”

worst aspect of vibe coding on a remote machine is u cannot drop a screenshot because it drops as a path on the current machine and the llm is like wtf are you talking about.. anyone got a fix?

@MajorBaguette might be related x.com/trq212/status/…

Meh.

openclaw is just the first way normies discovered AI agents but’s it basically Claude code with a crown scheduler and a “heartbeat” Claude code initiated most agentic things as we know that so makes sense they’ll keep pushing and innovating toward more and more autonomous agentic system

they're building openclaw

@noahzweben omg YES thanks 😈

Announcing a new Claude Code feature: Remote Control. It's rolling out now to Max users in research preview. Try it with /remote-control Start local sessions from the terminal, then continue them from your phone. Take a walk, see the sun, walk your dog without losing your flow.