Good morning! This is your daily ☕️ Techpresso.
In today's Techpresso:
🤖 Google is working on reasoning AI
👀 OpenAI co-founder joins rival Anthropic
🗣️ OpenAI introduces new multimodal processing, AI fine-tuning tools at DevDay
📱 Apple's next iPhone SE leaks
🫠 Snap employees were well aware of the app’s child safety issues
🤗 Telegram CEO Pavel Durov insists ‘little has changed’ in app’s privacy policies since his arrest
🎁 + 7 other news you might like
🔮 + 1 handpicked research papers and tools
🤖 Google is working on reasoning AILINK
Google is creating a new artificial intelligence model to compete with OpenAI's o1, focusing on logical reasoning and complex problem-solving skills.
Reports indicate Google teams have advanced on the "AI reasoning software," employing a "chain of thought" approach to tackle multi-step problems by generating, evaluating, and selecting the best answers.
GoogleDeepmind's research explored using additional computing power during inference to enhance model efficiency, resulting in more than four times the improvement in performance compared to traditional methods.
👀 OpenAI co-founder joins rival AnthropicLINK
Diederik "Durk" Kingma, one of OpenAI's co-founders, has announced his move to Anthropic, marking another high-profile departure from OpenAI to this rival AI firm.
Kingma, renowned for his influential work like the Adam optimizer and Variational Autoencoders, will join Anthropic after his recent stint at Google, adding significant expertise to their team.
Kingma will primarily work remotely from the Netherlands but plans to visit Anthropic's Bay Area offices occasionally, aligning with his commitment to their responsible AI development mission.
🗣️ OpenAI introduces new multimodal processing, AI fine-tuning tools at DevDayLINK
At the 2024 DevDay event, OpenAI revealed the public beta release of its new Realtime API, which enables developers to create advanced voice assistants using a single API call.
OpenAI plans to enhance its Chat Completions API by adding audio input and output capabilities, allowing developers to receive responses in text or audio format.
To help balance performance and cost, OpenAI introduced "model distillation" and "prompt caching" features, enabling efficient and cheaper AI application development.
📱 Apple's next iPhone SE leaksLINK
The next iPhone SE, resembling the iPhone 14, is expected early next year and will feature Face ID and Apple Intelligence capabilities, according to Bloomberg’s Mark Gurman.
The new iPhone SE models will likely ditch the home button and chunky bezels, potentially adding an OLED screen instead of the usual LCD, indicating significant design and performance upgrades.
Apple will also launch 11-inch and 13-inch iPad Airs with internal improvements alongside new Magic Keyboards and M4-equipped Mac Mini, MacBook Pros, and iMacs before the year ends, Gurman reports.
🫠 Snap employees were well aware of the app’s child safety issuesLINK
By November 2022, Snap employees were aware of about 10,000 monthly user reports of sextortion, acknowledging that these reports likely represented only a small fraction of the actual abuse occurring on the platform.
Internal research in 2022 allegedly found that over one-third of teen girls and 30% of teen boys on Snapchat had been exposed to unwanted contact, indicating significant safety issues for minors using the app.
The unsealed complaint claims Snap employees worried that measures against child predators and illegal activities, such as the sale of guns, would burden user privacy and administrative costs, and faced pushback due to CEO Evan Spiegel's prioritization of app design.
🤗 Telegram CEO Pavel Durov insists ‘little has changed’ in app’s privacy policies since his arrestLINK
Telegram CEO Pavel Durov asserted that the app's privacy policies have remained largely unchanged since his arrest, emphasizing that the platform has been sharing user data with authorities for years to combat criminal activities.
Durov clarified that since 2018, Telegram has provided IP addresses and phone numbers of alleged criminals to government authorities, complying with properly formed legal requests.
Facing accusations in France of facilitating criminal activity through lax content moderation, Durov defended himself by stating that Telegram’s core principles and commitment to user privacy and freedom have always been maintained.
Other news you might like
Microsoft exec reassures staff there’s no plan for Amazon-style RTO.LINK
Amazon recently rolled out a new AI chatbot that is 'safer than ChatGPT' for employees to use.LINK
Nintendo has reportedly shut down Ryujinx, the Switch emulator that was supposedly immune.LINK
Ubisoft investors push for company sale as shares hit decade-low.LINK
Automattic demanded a cut of WP Engine’s revenue before starting WordPress battle.LINK
Christian Selig’s unofficial YouTube app for the Vision Pro just got taken down.LINK
Pika 1.5 launches with physics-defying AI special effects.LINK
Latest research and tools
ScanOSS: it identifies and reports profanity usage within open-source software projects.LINK
Want to get the latest news differently? Find us on:
See you tomorrow for a new dose of ☕️ Techpresso!
-min.png){kind=small}
.png){kind=small}
