OpenAI’s New Move

but there's a catch!

Welcome to AI Wire — your smart shortcut to all things AI, without the jargon.

What we’ll cover today:

🤑 OpenAI cuts prices in half (but there's a catch)

🔐 Ex-YC Prez launches $100K fund for Safe AI

🥊 Chatbot Arena becomes a company — here’s what’s changing

🛠️ Gemini 2.5 Flash lets you toggle AI thinking like a switch

Source : PYMNTS

OpenAI (the ChatGPT people) launched “Flex processing” — a slower but cheaper way to use their AI models, o3 and o4-mini.

🚨 Breaking News
Flex lets users save 50% on API costs, but it runs slower and sometimes might not be available. Trade-off central.

⚡ Wire Simplified

  • Flex = 50% off if you’re okay with slower replies and random delays.

  • Great for non-urgent stuff like testing, research, or data crunching.

  • o3 model drops to $5 input, $20 output per million tokens (vs. $10/$40).

  • o4-mini becomes dirt cheap: $0.55 input, $2.20 output.

  • Also, OpenAI’s adding ID checks for some users to stop misuse.

✔️ Straight to the Point
OpenAI’s Flex mode makes AI way more affordable—but you’ll need to be patient. Great move for devs, annoying for real-time users.

⚖️ Would you trade speed for lower cost when using AI?

Login or Subscribe to participate in polls.

Source : Y Combinator

Geoff Ralston, ex-president of Y Combinator, launched SAIF — a fund investing $100K in startups making AI safer and more ethical.

🚨 Breaking News
SAIF backs early AI safety startups and gives them YC-style mentorship. It's a “safe” fund — literally and figuratively.

⚡ Wire Simplified

  • SAIF = Safe Artificial Intelligence Fund (yup, pun intended).

  • $100K per startup via SAFE agreements, capped at $10M valuation.

  • Focus: AI safety, IP protection, misinformation tools, and ethical governance.

  • Ralston also offers coaching, YC application tips, and weekly office hours.

  • Think of it like YC-lite… but only for “good” AI.

✔️ Straight to the Point
SAIF is Ralston’s bet that startups solving AI risks are the future—and they might make money while saving humanity too.

🧠 Should more VCs focus on safe AI instead of just fast-growing ones?

Login or Subscribe to participate in polls.

Source : Hugging Face

UC Berkeley’s viral AI model-testing site, Chatbot Arena, is now a real company: Arena Intelligence Inc., branded as LMArena.

🚨 Breaking News
LMArena will stay neutral but now plans upgrades, new features, and maybe even charges for model testing. Big moves ahead.

⚡ Wire Simplified

  • Born at UC Berkeley, now a legit company run by its original researchers.

  • One million+ users compare AI models head-to-head—like a rap battle but for bots.

  • New site beta.lmarena.ai is faster and mobile-friendly.

  • Still no business model, but may charge AI companies to test models.

  • Community trust and neutrality stay top priority (no brand favoritism).

✔️ Straight to the Point
LMArena is leveling up from a nerdy side project to a real business—just hoping it keeps its honest, no-bias roots.

🛠️ TOOLS

📚 RESOURCES

This 3-hour crash course teaches you what AI agents are, how to build them step-by-step, and how to make money from them—no tech background needed.
Perfect for anyone who wants to tap into the booming AI market without writing code.

👉 Watch the full beginner’s guide here – it might just be your shortcut to a new income stream 💸

Source : 9GAG

Was this forwarded to you? Sign up here.

AI Wire News. 

Signing off