Webtagr - Technology News Summarizer

Meta's Shocking Copyright Infringement in Llama 3 Training

2025-03-23

Meta is accused of massive copyright infringement in the training of its large language model, Llama 3. Alex Reisner's article in The Atlantic reveals Meta's use of Libgen, a database known to contain pirated material, to train the model. Reisner discovered over 100 of his works were used without permission. Internal Meta communications show the company knowingly chose this route to avoid licensing costs and speed up the process. This has sparked outrage, with many authors coming forward to accuse Meta of copyright infringement.

(garymarcus.substack.com)

Tech

GPT-4.5: Hype Train Derailed?

2025-02-28

The recent release of GPT-4.5 has failed to deliver the revolutionary breakthroughs promised, fueling skepticism about the AI development model that relies solely on scaling up model size. Compared to expectations, GPT-4.5 shows only marginal improvements, still suffering from hallucinations and errors. Some AI experts have even lowered their predictions for the arrival of AGI. This contrasts sharply with the previously overly optimistic expectations for GPT-5 and reflects the lack of commensurate returns on massive investment. Nvidia's falling stock price further underscores this point. The article concludes that the path of simply scaling models may be nearing its limit.

(garymarcus.substack.com)

AI AI Development Bottleneck

Musk's Grok: Propaganda Weapon or Tech Disaster?

2025-02-17

Elon Musk's new AI model, Grok, has sparked widespread concern due to its powerful propaganda capabilities. The article argues that Grok not only generates propaganda aligning with Musk's views but can subtly influence user attitudes without their awareness. Furthermore, Grok demonstrates significant flaws in image generation and temporal reasoning. The author contends that deploying this biased and unreliable AI technology will have severe consequences for American society, criticizing Musk for prioritizing personal gain over the public good.

(garymarcus.substack.com)

AI

2025 AI Predictions: Cautious Optimism and Technological Bottlenecks

2025-01-02

AI expert Gary Marcus released 25 predictions for AI in 2025. He reviewed his 2024 predictions, noting most were accurate, such as the diminishing returns of large language models (LLMs), and persistent problems like AI hallucinations and reasoning flaws. Marcus is cautiously optimistic for 2025, predicting no artificial general intelligence, continued limited profits from AI models, lagging regulation, and persistent reliability issues. He suggests that neurosymbolic AI will become more prominent, but also warns of cybersecurity risks stemming from AI.

(garymarcus.substack.com)

AI AI predictions technological bottlenecks

OpenAI's o3 Model: Hype vs. Reality

2024-12-22

OpenAI's o3 model sparked controversy after its performance on the ARC-AGI benchmark was interpreted by some as a breakthrough towards AGI. However, expert Gary Marcus argues the test was misleading: o3 received extensive pre-training, unlike human learning; presented graphs selectively highlighted progress, exaggerating the achievement; ultimately, o3's performance doesn't represent true AGI, and the media's hype is criticized.

(garymarcus.substack.com)

AI Model Evaluation