2025 has already brought us the most performant AI ever: What can we do with these supercharged capabilities (and what’s next)?

The latest AI large language model (LLM), for example, is releases from Xai to anthropic and grock 3 in Klod 3.7 frequently performed At PhD levels – at least due to certain criteria. This success records the next step at Google CEO Eric Schmidt irritate: A world with access to a “big polymat” is an EU that can lead to a wide range of knowledge to solve complex problems between subjects.

Wharton Work School Prof. Ethan Mollick also noted Something useful This recent models have used more than the power of computing more than GPT-4, and up to 10 times more than 10 times more than GPT-4. He added that this “this new generation is smarter and will make the Gen 3” AI model of the first “first” AI model “first” the first “AI model.

For example, Claude 3.7 shows the ability to review new angles in user needs and problem solving. According to anthropic, the first hybrid The justification model connecting a traditional LLM for quick answers with advanced thinking skills to solve complex problems.

Mollick belonged to two converting trends in these progress, these two trends closed that the two trends are “Super Loading AI Abilities”.

What can we do with this super-loaded AI?

In an important step, Openai has initiated “Deep Research” AI Agent in early February. In his opinion PlatformerCasey Newton said that the depth investigations appeared “effectively skilled.” Newton, deep research and similar tools can significantly accelerate research, analysis and other knowledge work, although the reliability in complex fields is still an open question.

Based on an option of still unacceptable O3 Reasoning model, deep research can deal with extended justification over a long time. It uses a very logical steps to divide the complex tasks in many logical steps as a human researcher, a thoughtful and thoughtful (COT) reasoning. You can also search the Internet to get more relevant information than the model training information.

Timothy wrote Lee Understand AI Deep research on several test experts said, “His performance has shown the effective capacity of the main o3 model,” he said. A test asked how to build a hydrogen electrolysis plant. Interpreting the quality of the speech, a mechanical engineer “estimated that there would be an experienced professional in a week 4000 word report Openai was formed in four minutes. ”

But wait, the more …

Google Deprmind also recently released “AI scientist, the multiplied AI system based on the Gemini 2.0 LLM. The novel to scientists is designed to help create assumptions and research plans. Now, the Imperial College London proved the value of this vehicle. Professor José R. Penadés, according to his words The team spent years Why do certain superbugs prevent antibiotics. AI repeated his findings in a total of 48 hours. AI sharpened assumptions, the generation of hypothesis, human scientists were still needed to confirm the findings. Nevertheless, Penades gossip The new AI application has a “Super boot potential.”

What can i mean for Supercare Science?

Last October, anthropic CEO Dario Amodi wrote in him “Loving grace machines“His waiting blog” strong AI “- the most rich artificial intelligence (AGI) called the most rich (AGI) – from the next 50 to 100 biological [research] Progress for 5 to 10 years. “Four months ago, the idea of ​​squeezing a century of scientific progress was extremely optimistic. Now the latest advances in the Openai Deep Research and Google EU, including AI models, which are called the latest advances, a nearby” radical transformation “begin to look more important.

However, the AI, scientific discovery, biology, at least, can still be limited to experimental assessment, regulation confirmation and clinical trials – real world restrictions. The question is already changing the science of the AI ​​(of course, which will be), on the contrary, the fully affects the fully.

On February 9 macal Post, Openai CEO Director-General Sam Altman “Systems that start to point to AGI, the systems that begin to appear.” AGI described the “increasingly complex problems, as a system that can solve the human level in many areas.

Altman believes that this stage is achieving, “We look amazing and now we can imagine a world that we treat all the diseases and imagine the world that can fully implement our creative potential.”

Modesty dose

These developments of AI have a very important and very different future in a short time. Again, the AI’s meteoric rise is not stumble. Humane Ai Pin review the last collapse – a device like a smartphone change after a noisy Tedi. A year later the company collapsed and this The residues were sold once for a part of high valuations.

Real-World AI applications are facing significant obstacles for many reasons for the infrastructure limitations for many reasons. This was undoubtedly a beginner Sensei AG experience supported by one of the richest investors in the world. The company left the AI ​​to apply the AI ​​to agriculture, using robots to produce a range of products to agriculture, but got acquainted with great obstacles. According to The Wall Street Journal has faced many failures to the unexpected logistics, stressing the gap between the Potential and practical execution of the EU’s potential and practical execution.

What’s next?

As we looked at the near future, science is due to the fact that a new gold discovery period is an increasingly capable partner in the field of research. Deep learning algorithms operating in Tandem with a human interest, AI systems, an information, spot samples, sensitive disciplinary assumptions.

Scientists already use a compression to constrict research time – reduction of protein structures and reducing literature until months or even days – opening opportunities in the medical areas of climate.

However, the potential of radical transformation is becoming clearer, so do the risks of disruption and instability. Altman himself, “The power balance between capital and labor can be easily involved,” EU’s economic impact was a delicate but significant warning.

This concern is already in Hong Kong, as soon as the city has been demonstrated Cut 10,000 Civil service work while walking AI in the same time. If such trends continue and expanded, we saw a widespread workforce, increasing social unrest and increasing social disorders and put pressure on the world and governments and governments.

Adaptation to the world of AI-Gearted

The possibility of increasing a deep change in the scientific discovery, justification and decision making of AI, but an extraordinary promise and giant problems. The forward road can be celebrated with economic violations and institutional strains, history and societies are indicated that it can always be able to adapt to technological revolutions, not always easily or ineffective.

To successfully manage this transformation, societies must invest in adaptation to management, education and labor to ensure equal distribution of the EU benefits. As AI regulation, political resistance, scientists, politicians and work leaders, ethical frameworks should cooperate, facilitating the transparency standards and crafts policy in strengthening the transformative effects of the EU. If you have raised this problem with this problem with Faraght and responsibility, people and AI can solve the world’s greatest difficulties in a new age with impossible improvements.