haroldo

dev and AI enthusiast.

[haroldo @haroldodev]

@themishra4402 you're crazy lmao

[haroldo @haroldodev]

@LordoftheMounts 8 min running, retry after retry, and I've already hit my quota. Antigravity didn't touch ONE SINGLE FILE. There's definitely something wrong.

[haroldo @haroldodev]

Opus 5.5 on Antigravity...

[haroldo @haroldodev]

Meanwhile, OpenAI is getting cooked by everyone because of their new plans, problems with Dots, and promised resets that never has been shipped.

Seems like Anthropic is nailing right now.

[haroldo @haroldodev]

Gemini 4 Argon is leading AutomationBench-AA, from Artificial Analysis. What a launch... It's a shame that it is a private launch. Release the dragon, Google.

[haroldo @haroldodev]

Gemini 4 Argon is leading AutomationBench-AA, from Artificial Analysis. What a launch... It's a shame that it is a private launch. Release the dragon, Google.

[haroldo @haroldodev]

@noclipepe fake

[haroldo @haroldodev]

@Priyannkaaaa Google. AI isn’t everything.

[haroldo @haroldodev]

@shengzheyao you guys should focus on antigravity desktop, not the CLI. i would bet that >80% of users would rather use antigravity desktop than antigravity cli.

[haroldo @haroldodev]

Google needs a new @antigravity desktop asap. The current Antigravity is pure trash when compared to Claude Code, Codex, ZCode and even OpenCode.

[haroldo @haroldodev]

I used to like this guy and his reviews, but he just sounds like a full-time OpenAI hater right now.

[BridgeMind @bridgemindai]

I ran out of my entire weekly limit on the $500 ChatGPT Pro Max plan in under 30 minutes.

GPT 6 Astra Ultrafast.

0% left until October 6.

OpenAI just scammed everybody.

[haroldo @haroldodev]

@ParamSiddh for now, it’s only for ultra users, but they will rollout asap. at least that’s what they said

[haroldo @haroldodev]

@AbdoKerdawy why are u screaming????

[haroldo @haroldodev]

in just two weeks, my feed went from being full of Codex fans to Codex haters. is openAI cooked?

[haroldo @haroldodev]

@MiaAI_lab @GeminiApp nice ragebait

[haroldo @haroldodev]

Benchmarks sometimes feel so fake. GLM 5.3/Flash is far superior to Muse Spark 1.3 on many tasks. I've used both A LOT, and GLM 5.3 has been much more accurate and consistently given me better results. Yet, according to AA, Muse Spark 1.3 is supposed to be better.

[haroldo @haroldodev]

@t_wassey tiktok

[haroldo @haroldodev]

@WillDobrev pq tenho o Pro de graça e não sou hater igual a maioria, mas uso mais pra tarefas simples e mecânicas nos modelos atuais disponíveis (até o 3.8 flash). to bem ansioso pra testar o Gemini 4 Argon

[haroldo @haroldodev]

@devcelio se vc não sabe o que cada linha do código faz, você não deveria nem ser contratado a depender da posição.

[haroldo @haroldodev]

@alwayspriyesh that said…

https://t.co/kX9IcLqE3a

[Google DeepMind @GoogleDeepMind]

Introducing Gemini 4 Argon – our new frontier model.

It’s built for complex workflows across coding, enterprise knowledge work, and cybersecurity defense – rolling out today to a set of trusted testers through our Fairwind Program.

[haroldo @haroldodev]

if Gemini 4 is that good, I won't renew my OpenAI and Claude subscriptions any soon. GOOGLE IS SO BACK