Scrolling through Twi tonight, I saw a blogger I follow recommending this AI tool called @Surf_Copilot. I decided to give it a try with a 'why not' attitude. Just a few days ago, I was complaining about how GPT and Grok couldn’t even calculate the difference between coin-margined and USD-margined liquidation prices. I was curious to see how this crypto-specific Copilot would perform. First, I tested it with the liquidation price question that GPT and Grok both got wrong a few days ago. To double-check, I went back and asked GPT and Grok again. This time, GPT got it right, Grok gave a quick wrong answer but managed to get it correct after a few minutes of "deep thinking." Surf’s performance was similar to Grok’s—quick response was wrong, but after deeper analysis, it got it right. Then I asked about the investment prospects of kaito, which has been trending tonight. Now, this answer was pretty interesting—it actually made a lot of sense! Especially with the detailed breakdown: the airdrops already launched, Mira’s earnings today, potential future airdrops, the subsequent selling pressure, and so on. All the data was laid out clearly. For this question, Surf’s answer was way better than what Grok and GPT provided. As for the price prediction, Surf suggested a range of 1.5-1.8, but recommended cashing out by the end of this year. It pointed out that the unlocking starts in February next year, and the dilution over the year would be significant. What’s unclear right now is whether Surf is using some open-source large model fine-tuned with crypto data, or if it’s pre-trained directly on crypto-specific data to create its own vertical large model. There’s too little public information available. But purely from a crypto research user experience perspective, it’s pretty solid. If you’re interested, you can give it a try. Here’s my invite link: https://asksurf.ai/?r=79LPO1SY620D