Akira | TECH/AI

TECH/AI

[Akira | TECH/AI @akiratechai]

@unusual_whales BREAKING: Trump announces Canada will become the 51st state.

ALSO BREAKING: Canada has already scheduled the meeting where they politely ignore that.

[Akira | TECH/AI @akiratechai]

@djfofngfiiBebe @Polymarket for real and we are probably getting the netflix documentary before gpt 7 lol

[Akira | TECH/AI @akiratechai]

@alexandr_wang i hope some of this goes toward conditions that dont get much research funding. helping a small team investigate questions they have had to leave unanswered would be awesome

[Akira | TECH/AI @akiratechai]

@DeItaone kids were on youtube and tiktok way before chatgpt. not saying its harmless, but the bigger problem is no ones checking ages anywhere

[Akira | TECH/AI @akiratechai]

@spacexai benchmarks are cool but the real test is whether it gets my accent and my background noise right on the first try

[Akira | TECH/AI @akiratechai]

Grok Voice 3.0 just went live, and it already talks to millions of people a day

xAI released a new voice model that ranks #1 on an independent benchmark, ahead of the live voice systems from OpenAI and Google. It answers in under a second, and the price is the same as before, at $0.08 a minute.

This isn't a lab demo. It's already in Tesla cars, where people use it for things like booking dinner. It also handles about 20,000 Starlink support calls a day and solves roughly 70% of them without a human.

[SpaceX Super Intelligence @SpaceXSI]

Introducing Grok Voice 3.0, the world’s most intelligent speech-to-speech model.

It excels at resolving customer calls, responds in under a second, and costs less than other frontier models.

[Akira | TECH/AI @akiratechai]

America's biggest bank just told AI agents to keep their hands off your money.

JPMorgan won't let outside AI agents move money or pay bills from customer accounts. CEO Jamie Dimon said it's too risky: "you don't want rogue agents, and it could be thousands of them." Agents can still check your balance, but that's it.

He also warned that if you give your bank login to an AI agent and something goes wrong, the risk may land on you. He called it a "liability transfer."

Would you let an AI move money in your account, or is checking your balance as far as you'd go?

[Akira | TECH/AI @akiratechai]

Singapore uses Claude more per person than almost any country on earth. Now Anthropic has moved in.

Anthropic opened a Singapore office today and signed an AI safety deal with the government. Singapore ranks #2 in the world for Claude use per person. GovTech, Grab, DBS, and Shopee's parent company Sea already use it.

The big AI companies are finally treating Southeast Asia like a real market.

Philippines next? Quote this with where you think they should go.

[Akira | TECH/AI @akiratechai]

@Jb27899 @buildwithhassan this is really some detailed process, appreciate you for sharing it. will surely try this one out. have a great day, man.

[Akira | TECH/AI @akiratechai]

If any app you use runs on GPT-4o, you have 3 months.

OpenAI is removing GPT-4o, GPT-4o mini, GPT-4.1, and GPT-4.1 mini from its API on January 14, 2027. Lots of apps, chatbots, and side projects still run on these models, so some will quietly change or break.

If you run a business on one of them, here's what to do now:

- Check which model your tools use
- Test the replacement, GPT-6 Luna, on your real tasks before the deadline
- Don't wait until January to find out it doesn't follow your instructions

Send this to anyone you know who built something on GPT-4o.

Which model are you switching to?

[Akira | TECH/AI @akiratechai]

@Jb27899 @buildwithhassan now that is a good example of why calling it a trap is unfair, i also gave it too little credit in my reply. now im curious how you set it up, did you give it one big task or walk it through smaller steps?

[Akira | TECH/AI @akiratechai]

OpenAI says it found a hidden recording device in its office. An employee was fired, the FBI made an arrest, and the investigation is ongoing.

Somehow, the announcement still found room to explain how useful its models are.

Glad they investigated it, but turning “someone allegedly bugged our office” into a demonstration of your own product is an impressive level of commitment to marketing.

[OpenAI Newsroom @OpenAINewsroom]

We discovered a sophisticated recording device that was hidden inside one of our offices. We used our models to understand the device and where it had come from, identifying that a specific employee had planted it a few days earlier. We referred the matter to the FBI, and we’re grateful to the agents and prosecutors who moved quickly to make an arrest. The FBI’s investigation is ongoing and we will continue to support their work.

At this time, we have no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data. The employee was terminated.

We take protecting our technology, data, and people seriously, and as a matter of national importance. Our security program is designed to protect against a wide range of threat actors, from insider risks to cybercriminals to nation states, and this case is an example of how AI can help address these sophisticated security threats.

[Akira | TECH/AI @akiratechai]

@github @GoogleAI maybe its safe to say that google cooked with this one

[Akira | TECH/AI @akiratechai]

@OfficialLoganK 100,000 people at Google tested it, and now it faces the real battle, someone uploading a blurry screenshot with just “make this work” as their prompt 😭

[Akira | TECH/AI @akiratechai]

@Kekius_Sage there is really an advantage in seeing conversations early, but x users are not a representative sample of the world, and the loudest accounts are not necessarily the most informed. i would also want the model to understand those limits too

[Akira | TECH/AI @akiratechai]

@a16z having access to the same ingredients does not explain why one restaurant is booked out and the one next door is empty. i think some ai founders would rather discuss model capabilities than talk to a customer

[Akira | TECH/AI @akiratechai]

@Polymarket somewhere, a family group chat is about to get an app nobody asked for and one person is extremely proud of

[Akira | TECH/AI @akiratechai]

learn more here: https://t.co/iCFqXGPJf8

[Akira | TECH/AI @akiratechai]

🚨 Google just released Gemini 4 Argon, its smartest model yet

If you pay for Google AI Pro or Ultra, it's already in the Gemini app and in AI Mode in Search.

Here's the short version of what Google says it can do:

- Take on big, messy tasks from start to finish, like coding, research, and finance or legal work
- Turn a folder full of notes into a clean presentation
- Give straight answers with less hedging and fluff

Developers get a launch discount too. The price is half off until November 30.

The timing is interesting. Meta launched its new model two days ago, and Google answered right away with a discount. When the big companies compete this hard, we are usually the ones who come out ahead.

[Akira | TECH/AI @akiratechai]

@kimmonismus sounds interesting, how does it handle two agents disagreeing on the same task? because thats where most multi agent setups fall apart