Andrew Altair

Insights

Short analysis and commentary from the AI world

ქართულად →
All#ai#anthropic#big-tech#openai#meta#google#ai-agents#claude#autonomous-systems#tech-industry#api#developer-tools#ai-regulation#prompt-engineering#gpt
Ling 3.0 Flash is the smartest open model at its size
the-decoder.com

Ant Group's Ling 3.0 Flash scores 38 on AI Index with hallucination rate dropping from 97% to 44%

Ant Group's inclusionAI laboratory released Ling 3.0 Flash, an open-weight language model designed for high efficiency in smaller parameter categories. On the Artificial Analysis Intelligence Index, the model achieved 38 points, matching Qwen3.6 27B while utilizing fewer active parameters. Evaluators identified it as the highest-scoring open model under 124 billion total parameters. In reliability testing on the AA Omniscience benchmark, the model's hallucination rate decreased from 97% to 44% compared to its previous version. The model is distributed under an MIT license on the inclusionAI API and DeepInfra, with open model weights hosted on Hugging Face.

#hugging-face#open-source-ai#api
Ling 3.0 Flash is the smartest open model at its size
Aug 19, 20262
Deepseek ships improved V4 Pro, open-sources its agent software, and raises API prices
the-decoder.com

DeepSeek ships V4 Pro update, open-sources its Harness v0.1 agent software, and raises peak API prices

DeepSeek released build V4-Pro-0813 for its flagship model, published its agent software as open source under the MIT license, and introduced new API pricing tiers. Benchmark results showed Terminal Bench 2.1 scores rising from 72.1 to 87.9, while DeepSWE coding evaluation jumped from 51.5 to 70.3. Artificial Analysis recorded the model's intelligence index score rising from 45 to 53 points. The newly released DeepSeek Harness v0.1 provides autonomous models with bash terminal access and file system navigation tools for automated coding workflows. Starting August 16, peak-hour API rates for 1,000,000 cached input tokens increased from $0.27 to $0.55, with output tokens priced at $3.30, while off-peak rates remained unchanged at $0.14.

#ai-agents#ai#ai-tokens
Deepseek ships improved V4 Pro, open-sources its agent software, and raises API prices
Aug 19, 20260
An AI agent for all? Try using your brain, Mark Zuckerberg | Brief letters
theguardian.com

British readers push back against Zuckerberg's superintelligent AI agents for replacing human critical thinking

Mark Zuckerberg's August 10 announcement pledging superintelligent personal AI agents for all users has drawn sharp public criticism in the UK. Letters published in The Guardian on August 13 raised concerns about delegating human cognition and decision-making to corporate AI platforms, warning of risks to education and youth employment. Contributors, including 16-year-old student Emad Rehman and researcher Michael Fuller, argued that relying on automated systems risks degrading independent thinking and critical evaluation skills among youth. Critics also noted that personal AI assistants acting as personal diaries create deep dependency on closed commercial cloud infrastructure.

#zuckerberg#meta#ai-agents
An AI agent for all? Try using your brain, Mark Zuckerberg | Brief letters
Aug 19, 20263
Mark Zuckerberg says the future of AI is for everyone. But who owns it? | Raffi Krikorian
theguardian.com

Mozilla CTO: Zuckerberg's "AI for all" vision leaves users renting intelligence without owning it

In an analysis published in The Guardian, Mozilla CTO Raffi Krikorian responded to Mark Zuckerberg's 6,500-word essay "The Future Is for Everyone," arguing that closed corporate platforms do not provide true digital independence. Krikorian outlined three critical questions for AI users regarding platform lock-in, recalling Amazon's 2009 remote deletion of books from Kindles and the US government's shutdown of Anthropic's model on June 12, 2026. He emphasized that open-weight models allow organizations to download and run software locally without fear of external disconnection or data loss. Krikorian urged governments and critical institutions, such as healthcare providers, to maintain legally usable copies of open-weight models to ensure computational sovereignty and data portability.

#anthropic#ai#zuckerberg
Mark Zuckerberg says the future of AI is for everyone. But who owns it? | Raffi Krikorian
Aug 19, 20260
Facebook Is Filling With Viral AI-Generated Silver Foxes Wooing Women by Ranting About How Terrible Men Are
futurism.com

Viral AI-generated male podcasters are racking up millions of views on Facebook by pandering to women

A new wave of AI-generated video clips featuring synthetic male podcast hosts lecturing women about relationships is spreading across Facebook, generating millions of views. Characters featured on pages like "The Male Insight" and "Leo Davidson" mimic podcast video clips, delivering scripted monologues about how modern men lack accountability. Videos on these pages frequently reach 200,000 to 280,000 views, with one clip surpassing 1,000,000 views. According to a Futurism investigation, the primary audience consists of older women who believe the AI hosts are real relationship experts. The creators monetize the synthetic content through Facebook ad revenue, AI-written e-books, and texting guides. The synthetic personas even fooled Google's AI Overview, which indexed the fictional character "Roman Black" as a real author and podcaster before corrections were applied.

#google#big-tech
Facebook Is Filling With Viral AI-Generated Silver Foxes Wooing Women by Ranting About How Terrible Men Are
Aug 19, 20261
College Kids Are Using AI Agents to "Take" Entire Online Classes for Them by Logging Directly Into the Course Management Software
futurism.com

College students deploy autonomous AI agents to take entire online courses and submit exams

An investigation reveals that university students are deploying autonomous AI browser agents to take entire online college courses. The agents log into learning management systems like Canvas and Blackboard, watch required video lectures, complete weekly quizzes, and write discussion board responses. Major AI models including ChatGPT, Gemini, and Grammarly readily execute academic assignments when prompted. Perplexity representatives stated that screening or blocking academic platforms would compromise user privacy and security, declining to restrict agent access. In response, institutions are reversing remote grading policies. The University of Chicago Law School banned laptops and phones from 1L lectures, while Princeton University ended a century-old unproctored exam honor code to curb automated cheating.

#openai#gpt#ai
College Kids Are Using AI Agents to "Take" Entire Online Classes for Them by Logging Directly Into the Course Management Software
Aug 18, 20261
Insane Dad Boasts About Using AI to Construct a "Truman Show" Style False Reality For His Son To Play In
futurism.com

Father creates fake Roblox reality populated entirely by ChatGPT bots to isolate 6yo son

Software developer Adam Lyttle created an isolated clone of Roblox for his 6-year-old son, populating every other character and chat participant with ChatGPT-driven AI bots. The child played under the assumption that he was interacting with real human peers online. The custom game featured simulated player sessions and integrated spelling mini-games in chat. When shared publicly, the project sparked fierce backlash, with critics drawing comparisons to the 1998 film 'The Truman Show' and warning against deceptive synthetic environments. The incident highlights growing ethical debates surrounding AI in parenting. OpenAI CEO Sam Altman recently disclosed using ChatGPT daily for child-rearing guidance, noting possibilities like personalized AI podcasts tailored to a child's day.

#openai#ai#gpt
Insane Dad Boasts About Using AI to Construct a "Truman Show" Style False Reality For His Son To Play In
Aug 18, 20262
Gemini 3.7 Flash lands with coding gains and undercuts its three-week-old predecessor's price by 50%
the-decoder.com

Google launches Gemini 3.7 Flash with coding gains and 50% price reduction

Google has released Gemini 3.7 Flash just three weeks after deploying Gemini 3.6 Flash, demonstrating an accelerating model update cadence. Designed specifically for software engineering and autonomous agent workflows, the model delivers notable benchmark gains alongside a 50% API price drop. On the FrontierCode benchmark, Gemini 3.7 Flash jumped from 34.4% to 43.6%, while its DeepSWE evaluation reached 65.3% (up from 49.0%), outperforming Claude Sonnet 5 and GPT-5.6 Terra on key software engineering metrics. Pricing is set at $0.75 per million input tokens and $3.75 per million output tokens, matching the reduced tier across Google AI Studio and Antigravity. The aggressive pricing strategy intensifies competition in the developer tool and agent framework markets.

#openai#gpt#ai
Gemini 3.7 Flash lands with coding gains and undercuts its three-week-old predecessor's price by 50%
Aug 18, 20260
Building a practical path to post-quantum cryptography
technologyreview.com

Quantum computers may crack 2048-bit RSA keys by 2040 as NIST rolls out post-quantum standards

A survey of 32 cryptography experts conducted by the Global Risk Institute reveals a 50% likelihood that quantum computers will break 2048-bit RSA encryption within 24 hours by 2040. Security researchers warn that the primary threat is 'harvest now, decrypt later' (HNDL), where adversaries capture encrypted traffic today to decrypt once quantum machines scale. In response, the US government issued CNSSP-15 requiring national security systems to adopt CNSA 2.0 post-quantum standards starting January 2027, with complete transition targeted for 2035. NIST has finalized standard post-quantum algorithms, pushing hardware and enterprise vendors to upgrade infrastructure. Hardware manufacturers are implementing protections at the silicon layer. Intel's Xeon 6 processors integrate AES-256 memory encryption and QuickAssist acceleration offloading to maintain high throughput during quantum-safe handshakes.

#crypto#web3
Building a practical path to post-quantum cryptography
Aug 18, 20260
Suno Studio 2.0's new chat feature lets you talk to your DAW like it's a bandmate
the-decoder.com

Suno releases Studio 2.0 with chat-driven DAW plugins and multitrack audio export

AI music generation platform Suno has officially rolled out Studio 2.0, adding an interactive text chat feature directly into its digital audio workstation (DAW). The update allows Premier tier subscribers to generate instruments, vocal lines, and custom audio plugins conversationally. According to Suno CEO Mikey Shulman, the goal is to eliminate routine technical hurdles in music production: "Our aim is to create an environment where natural language instructions turn directly into working studio tools." The platform now supports MIDI file import, stems separation, automation curves, and multitrack master export at 32-bit depth and 48 kHz sample rates. While plugin generation currently does not consume extra credits, Suno noted that usage limits may be introduced later. The move follows a German court ruling finding Suno liable for copyright infringement, reinforcing regulatory scrutiny across European markets.

#automation#ai#copyright
Suno Studio 2.0's new chat feature lets you talk to your DAW like it's a bandmate
Aug 18, 20261