Pinned Tweet
I've been running OpenClaw as my always-on personal AI assistant for months — Telegram, Discord, WhatsApp, iMessage, Gmail monitoring, calendar, proactive alerts, the works.
Claude Code Channels + Cowork + Computer Use have closed the gap dramatically in just one week. But here's what's still missing for a full replacement:
1. Daemon mode — OpenClaw runs as a headless LaunchAgent 24/7. Claude Desktop needs a GUI session open. Ship a claude-daemon and it's game over.
2. Always-listening channels — OpenClaw receives messages and responds proactively. Claude Channels require an active session with --channels flag. No event-driven "inbox".
3. WhatsApp/Signal — Channels only cover Telegram + Discord. Many of us live on WhatsApp.
4. Group chat participation — Channels are DM-only. Can't monitor and participate in Discord servers or Telegram groups.
5. Multi-provider routing — Sometimes you want GPT for a quick task or Gemini for a specific workload. OpenClaw routes across providers.
The trajectory is clear though. @trq212 and team shipped Dispatch, Channels, and Computer Use in ONE WEEK. At this pace, most of these gaps close within months.
The moment Claude ships a headless background service, the migration becomes real.
Object Orienter retweeted
Dos 14 candidatos, apenas dois são explicitamente bandidos.
Mesmo assim, 92% dos eleitores escolheram o crime.
A gente merece o Brasil que tem.
The biggest win from ChatGPT over Claude during these last 2 weeks I've been using ChatGPT is not having the very annoying 5 hour limit. I wish Claude had that.
19
Yep, for the first time in over a year I’ve cancelled my Claude subscription. Astra is really good.
GPT 6 Astra is so much better than Fable 5.1.
Fable 5.1 is slow, expensive, and lazy. I spent the past 12 months using Claude Code as my daily driver.
I use Codex now.
Astra is fast, reliable, and computer use is a genuine game changer. It does things no other model can do.
GPT 6 Astra has completely taken over my workflow.
205
Object Orienter retweeted
Not sure what he really wants to accomplish with this. Open models are not going to stop or slow down. We (as in Americans) would be handing off the leading position in AI to other nation states, in the best case scenario.
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: darioamodei.com/post/we-must…
Object Orienter retweeted
Most of the human code I've encountered is much worse than AI generated code from Fable and Astra
Tibo the true legend.
Never gonna give you up
Never gonna let you down
Never gonna run around and desert you
Never gonna make you cry
Never gonna say goodbye
Never gonna tell a lie and hurt you
Thanks for reading. We will do a global reset of the usage for all paid subscriptions so that you can keep enjoying Astra after burning through all of it doing fun 3D modeling in blender. The work week is about to start.
Lands around 6pm PST today.
64
Object Orienter retweeted
🚨URGENTE - Renan Santos diz que quer Flávio Bolsonaro preso e se recusa a participar das manifestações do dia 07/09 com ele
“Eu quero que Flávio seja preso! Se eu ganho a eleição eu vou prender o Flávio”
I got excited with GPT-6 and upgraded my plan from Plus to Pro x5... just to find out GPT-6 is still not available for me. Thanks @thsottiaux
53
Object Orienter retweeted
Talvez o maior escândalo dentro do caso Master seja o sumiço do celular de Alexandre de Moraes.
Não foi entregue.
Não foi apreendido.
Não foi periciado.
A sociedade está impedida de ver as respostas do magistrado da mais alta Corte do país ao banqueiro mafioso que - a julgar pelo conteúdo das mensagens apontadas em relatório policial - pedia ao amigo para exercer advocacia administrativa em seu favor ou obstruir a justiça, enquanto pagava mais de R$ 80 milhões ao escritório da esposa e dos filhos dele, referentes a um contrato de R$ 131 milhões com o banco.
De quebra, Daniel Vorcaro chegou a consultar Moraes, dois dias antes da prisão, sobre se deveria fugir do país: “Acha que segunda já tenho que estar fora?”
E não sabemos o que achou o ministro do STF, que continua dentro.
Ele é blindado pelo PGR amigo e por colegas que igualmente curtiram os eventos patrocinados por Vorcaro e a degustação de charutos e whisky Macallan.
Moraes nunca revelou o conteúdo das referidas conversas, não representou contra o banqueiro, não encaminhou o caso a autoridades para que as solicitações nada republicanas fossem investigadas.
Dois meses antes da prisão de Vorcaro, na verdade, o magistrado alterou a configuração no WhatsApp, fixando o prazo de 24 horas para a exclusão automática das mensagens trocadas com o banqueiro.
“Alexandre de Moraes usa uma duração padrão para mensagens temporárias em novas conversas. Todas as novas mensagens desaparecerão desta conversa após 24 horas”, diz um aviso do aplicativo, conforme print do chat de Vorcaro com o magistrado, presente no relatório da PF.
O próprio Moraes, ao votar pela condenação de Débora do Batom no caso do 8/1, apontou “desprezo para com as instituições republicanas” no “fato de que a ré apagou e ocultou provas de sua intensa participação nos atos golpistas”, “haja vista a conclusão apresentada pela Polícia Federal”, “relacionada ao celular de sua propriedade”.
Quando se compara o caso Master com o do jornalista do Maranhão, que teve notebook e celular apreendidos e sigilo da fonte violado, em busca e apreensão determinada por Moraes em razão de notícia verdadeira porém inconveniente ao colega Flávio Dino, nota-se também a dimensão da desigualdade entre um cidadão comum e um membro da casta, perante a oligarquia dos Poderes.
A casta não conseguiu bloquear a prisão de Vorcaro, mas bloqueou o acesso às respostas do marido de Viviane Barci de Moraes ao banqueiro. É escandaloso o que essa gente chama de “democracia”.
Object Orienter retweeted
Renan recuperou a campanha sem ameaçar botar bomba no STF, sem pedir intervenção militar, sem acampar em porta de quartel - o q faria perder apoio da imprensa. Manteve dureza e equilíbrio no discurso e, ao menos temporariamente, ganhou a queda de braço. O Brasil ainda não é uma ditadura.
Object Orienter retweeted
Outros candidatos não registraram no TSE perfis que usam em propaganda eleitoral.
Decisão de Toffoli de suspender campanha de Renan resultou na descoberta de casos até piores, porque o candidato do Missão, ao menos, fez o registro em 19/8, apenas 3 dias após o início oficial da corrida em 16/8.
Há os que não fizeram até hoje.
Fica a pergunta: todos terão a campanha suspensa? Ou só os que criticam Toffoli?
O correto, claro, seria o próprio ministro reconhecer a barbaridade que fez e revogar a decisão antidemocrática anterior.
Mas o correto não é o que ele costuma fazer.
Oooo Thomas, nao precisa tentar me convencer a votar no @RenanSantosMBL , eu já vou votar nele.
1
25
Crude coding, I agree, it does better than me. But the system architecture/engineering is off a lot of the times. Then I need to steer on those cases. But other than that, it has achieved 1. for me as well.
Replying to @mattpocockuk
tbh this is a good debate to have.
Here is what the timeline of the near past/future looks like to me:
1. Models are able to code better than I can
2. Models are able to do coding-adjacent engineering work better than I can: debugging, profiling, optimizing, bug fixing, system design, abstraction design, ui design, idea generation, etc.
3. Models are able to do most things that can be done on a computer better than most people
For the kinds of coding that I do, Claude has achieved 1 and some (but not all) of 2. It is showing early-ish signs of 3.
This is not the case for all coding yet. If you are Anders Hejlsberg and are a world class expert in compilers and type systems, Claude might not be at 1 yet. My bet would be that it gets there quite soon, while at the same time pushing further into 2-3 territory for the rest of us.
You can think of model capabilities as capturing the distribution of human ability. I consider myself an average programmer so Claude has already surpassed me, but the same is not yet true for everyone.
Back in Nov 2025 I fully stopped writing code by hand, though I continue to code (using agents) every day. That is the point where for me and many others around me, it felt fair to say that coding is solved. But engineering is more than coding. Claude’s code is not perfect. It has bugs and inefficiencies. Claude can code, but it cannot yet do everything that goes into engineering.
Opus 4.8 is the first model that felt like it could code better than me. Fable now routinely finds optimizations and debugs issues that I would not have been able to fix myself. And with each generation, the code the model produces continues to improve, and the model is able to do more of the non-coding parts of engineering.
21




















