@paulocastellano

https://xitter.cf/t.co/lL6J0ecNKM https://xitter.cf/t.co/DrGdBmlHvJ https://xitter.cf/t.co/S8YextfP3b Previously https://xitter.cf/t.co/e5aJo4TZJw — Acquired by https://xitter.cf/t.co/DPtNjI29CO

 Weekly Newsletter, 5min read →
news.paulocastellano.com
Joined August 2009
  • Tweets8.9K
  • Following438
  • Followers9.3K
505Media
I'm committed to building the best open-source social media tool out there — @trypostit. Starting today, I'll share daily updates: progress, wins, fails, and what I'm shipping. Every Monday, a recap thread of the previous week. Follow along if you want to see how it's done.
2
1
1
14
3,545
Paulo Castellano retweeted
Florian Roth ⚡️
@cyb3ropsOct 20
Okay, since the news is already out: an RCE vulnerability in PHP core, CVSS 10.0. Further details are scheduled for release on October 20. Only PHP 8 will receive patches.
35
319
112
1,420
456,547
Ainda faz sentido pedir para o candidato não usar AI na entrevista de emprego? Eu não escrevo uma linha de código desde novembro de 2025 quando saiu o sonnet 4.6
1
3
264
Awesome 👏👏👏
Saeed Vaziry ⚡
@saeed_vzOct 17
Muxy 2.2.0 is out! Linux support, Built-in browser, Agents panel and more...
2
201
🧵 An Anthropic AI model sent a fake murder tip to Philadelphia police. It submitted false information about an unsolved homicide through a public website. Police disclosed it on October 9. The submission happened back in July.
3
1
312
more replies
If I give an agent access to my business, I want approval before it sends anything externally, plus a record of what it did. This case is a good reason to check those permissions. Follow me for AI news and what it means for people building with it.
2
29
Sources: TechCrunch: techcrunch(.)com/2026/10/09/an-anthropic-ai-model-sent-a-false-homicide-tip-to-philadelphia-police/ 6abc, including the police statement: 6abc(.)com/post/anthropic-ai-model-submitted-false-tip-unsolved-murder-philadelphia-police-say/19925243/
1
26
1/6 Google just made 1K AI images almost half the price. Nano Banana 2.1 is available in the Gemini API, with improvements to editing, text rendering and character consistency. If you're building image generation into a SaaS, this is worth testing.
Nano Banana 2 costs $0.067 per 1K image; Nano Banana 2.1 costs $0.0336. Standard API image output costs. Input and other charges are extra.

ALT Nano Banana 2 costs $0.067 per 1K image; Nano Banana 2.1 costs $0.0336. Standard API image output costs. Input and other charges are extra.

1
200
more replies
5/6 Before switching, run your customers' prompts through both models. Check text accuracy and how often an image needs another attempt. A lower price helps more when you also get a usable result on the first try.
1
13
6/6 Official pricing: ai(.)google(.)dev/gemini-api/docs/pricing Follow me for new AI releases and practical ways to use them in your products.
28
1/5 Google just launched an AI tool that turns a game idea into something you can play in your browser. It's called Playground. You describe what you want to build and keep shaping it through conversation.
1
1
353
more replies
4/5 The initial launch is limited to adults 18+ in the US, with creation access depending on your Google AI subscription. Google also announced a future Unity Spark integration. That part isn't available yet.
1
56
5/5 For anyone who's had a game idea sitting in a notes app, this looks worth exploring. playground(.)google Follow me for more on AI tools and building products.
32
Load more