TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Show HN: Watch 3 AIs compete in real-time stock trading

270 点作者 sunnynagra5 个月前
A live dashboard where you can watch GPT-4, Claude 3, and Gemini analyze market data and make daily stock trades with real money. Each AI explains its reasoning, and you can compare their different approaches to the same data.<p>Link: <a href="https:&#x2F;&#x2F;trading.snagra.com?utm_source=hn" rel="nofollow">https:&#x2F;&#x2F;trading.snagra.com?utm_source=hn</a> (no signup required)<p>What you can try right now: - Watch live trades from GPT-4, Claude 3, and Gemini - Read each AI&#x27;s full analysis and reasoning - Compare their different interpretations of the same market data - Track their real-time performance and win rates - View historical trades and performance metrics<p>Built this over the holidays to study how different AI models approach financial decisions. Each morning at 9:30 AM EST, the AIs analyze market data and make real trades with $5 stakes.<p>Technical Implementation: - Next.js frontend with real-time updates - Node.js&#x2F;Lambda backend for AI processing - PostgreSQL for trade tracking - Alpaca API for automated trading - Consistent prompts for all models<p>Data Flow: 1. Daily market analysis (9:30 AM EST) 2. Each AI gets identical inputs: - Financial headlines - Market summaries - Technical indicators - Earnings reports 3. AIs provide: - Stock picks with reasoning - Entry&#x2F;exit conditions - Risk assessment 4. Automated trade execution<p>Note: This is an experiment in AI behavior, not investment advice. The goal is to study how different LLMs interpret financial data and make decisions with real consequences.<p>I&#x27;ll be around to answer questions about the implementation.

47 条评论

rixed5 个月前
&gt; The goal is to study how different LLMs interpret financial data and make decisions with real consequences.<p>I don&#x27;t really buy this. If the goal was to study how different LLMs interpret financial data there would be no use for actual trades, since their interpretation cannot be influenced by the fact that the trading orders are passed for real.<p>I believe the goal is to see if AI can do better than rats [0]. There is no shame in that.<p>[0]: <a href="https:&#x2F;&#x2F;www.vice.com&#x2F;en&#x2F;article&#x2F;rattraders-0000519-v21n12&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.vice.com&#x2F;en&#x2F;article&#x2F;rattraders-0000519-v21n12&#x2F;</a>
评论 #42560221 未加载
评论 #42560142 未加载
评论 #42560242 未加载
pakitan5 个月前
ChatGPT has one trade that is guaranteed to be bad. I&#x27;m not saying unprofitable, just bad. GBTC is the bitcoin ETF with biggest expense ratio - 1.5%. If you want to bet on bitcoin, a better choice would be BITB (0.20%) or BTC (0.15%).<p>Also, the reasoning is partially a hallucination - &quot;The holding period of 9 months aligns with the expected completion of Grayscale&#x27;s pivotal Phase 3 Bitcoin ETF trial, a major catalyst for unlocking investor demand and driving trust value realization.&quot;<p>There is no such thing as a &quot;holding period&quot;, nor are they doing a &quot;Phase 3 Bitcoin ETF trial&quot;. It&#x27;s possible the &quot;Phase 3&quot; thing is picked up from news about a drug company.
评论 #42561382 未加载
评论 #42592320 未加载
评论 #42562438 未加载
评论 #42562372 未加载
aredox5 个月前
They should have added a pure random bot as a control.<p>Or a monkey.
评论 #42560077 未加载
评论 #42562694 未加载
评论 #42560113 未加载
评论 #42560079 未加载
评论 #42560104 未加载
评论 #42563199 未加载
评论 #42560261 未加载
评论 #42563413 未加载
评论 #42573479 未加载
评论 #42561437 未加载
评论 #42561567 未加载
vasco5 个月前
&gt; Every morning at 5:45 AM PST, three AI models (GPT-4o, Gemini 1.5 Pro, and Claude 3 Sonnet) analyze the latest market news and each recommends one stock to trade.<p>&gt; At 6:00 AM PST, trades are automatically executed based on AI recommendations, investing $5 per trade<p>The best trading decision most days is to not trade. Outliers and diversions from the mean don&#x27;t happen every day. This is trading just for the sake of it.<p>I predict a slow crawl down into zero eaten up by fees.
NathanaelRea5 个月前
If they just get the financial headlines and indicators, aren&#x27;t they all just momentum trading from sentiment analysis?
评论 #42560065 未加载
评论 #42560399 未加载
评论 #42560088 未加载
评论 #42560067 未加载
AmazingTurtle5 个月前
Combining universal time-series prediction models with latent space global knowledge on realtime information could result in an accurate model prediction on the stockmarket with a bias towards succeeding. <a href="https:&#x2F;&#x2F;research.google&#x2F;blog&#x2F;a-decoder-only-foundation-model-for-time-series-forecasting&#x2F;" rel="nofollow">https:&#x2F;&#x2F;research.google&#x2F;blog&#x2F;a-decoder-only-foundation-model...</a>
clark-kent5 个月前
Very interesting idea. I&#x27;m thinking about creating an AI portfolio manager (private) that invests for the long term.<p>Some things to watch out for:<p>- LLMs, by default, don&#x27;t follow the best practices for trading or investing. Without careful constraints, they can ignore fundamental investment best practices. This is something I learned while building <a href="https:&#x2F;&#x2F;decodeinvesting.com&#x2F;chat" rel="nofollow">https:&#x2F;&#x2F;decodeinvesting.com&#x2F;chat</a>.<p>- I see Claude bought a penny stock SMX. This could be volatile, and the price could change significantly in 24 hours before the next execution at 9:30 am.<p>- The LLMs are day trading on some volatile securities; while LLMs could be good at day trading, unlike humans (we will find out), this setup has the disadvantage of only trading once a day.
评论 #42562356 未加载
wolfman15 个月前
Going to follow along to see how the results look in the months to come.<p>I&#x27;ve been working on the same concept for the past 2y now and have our performance results here: <a href="https:&#x2F;&#x2F;trend.fi&#x2F;performance" rel="nofollow">https:&#x2F;&#x2F;trend.fi&#x2F;performance</a>
评论 #42562448 未加载
评论 #42562230 未加载
detente185 个月前
Interesting — does your backend server use Python? I couldn&#x27;t find much about it on your site.<p>It would be great to see this tested with more commercial LLMs (O1 &#x2F; Amazon Nova, &#x2F; Llama 3.2 &#x2F; etc.). If you&#x27;re open to it, I’d be happy to contribute support for these models via LiteLLM - <a href="https:&#x2F;&#x2F;docs.litellm.ai&#x2F;docs&#x2F;providers" rel="nofollow">https:&#x2F;&#x2F;docs.litellm.ai&#x2F;docs&#x2F;providers</a>
aws-user5 个月前
Unfortunatly I can&#x27;t subscribe to the updates &quot;Failed to send verification email&quot;. Also, would you be willing to share what prompt are you using? Thanks!
评论 #42561946 未加载
noman-land5 个月前
Watch a random number generator generate random numbers.
评论 #42560120 未加载
johng5 个月前
My first email address it wouldn&#x27;t accept.. wouldn&#x27;t let me use it. Maybe the domain hit some censor (fscking.com)<p>Did a different email, it accepted it, I got the email, but got this error message when trying to confirm it: {&quot;error&quot;:&quot;Invalid verification token&quot;} and a pretty-print checkbox that did nothing.
评论 #42561947 未加载
geor9e5 个月前
I just asked ChatGPT 4o &quot;Guess what the average investor will do with todays stock market headlines. Just pick one specific trade.&quot; and it replied sell META. But your result was buy META. Could just be randomness, but I wonder if your prompt introduces a bias towards buying.
评论 #42560363 未加载
BadHumans5 个月前
Can I let Claude do all my trading for me? It currently sits at 77% unrealized gains.
lewj5 个月前
Is there any weighting towards selling in the negative? Else the LLM&#x27;s <i>should</i> just hold their unrealised losses, and only sell post local peak - depends on their suggested measurement of success?
评论 #42560147 未加载
评论 #42560107 未加载
attentionmech5 个月前
Related to this but little theoretical question - If you add an intelligent predictor of market which wins over other consistently by X% - then the market will start using that information and wouldn&#x27;t that make our intelligent predictor lose it&#x27;s edge?<p>More simply what i mean to ask is -&gt; the moment market knows about your advantage, shouldn&#x27;t you lose it because everyone else will use that information to balance the market?
评论 #42562319 未加载
评论 #42562330 未加载
jeffadelic5 个月前
How much are your infra costs for everything? And do you pay for the AI APIs or using free tier?<p>Really cool project and subscribed to follow along.
mvdtnz5 个月前
Mate your shitty app is sending tripled up email barrages. That is absolutely not ok and is illegal in many places.
ratedgene5 个月前
It would be neat to see the process, where they get the data from, how they analyze it.<p>It would be neat to also see another experiment of a MAS doing this and coordinating to gamble together. Perhaps even different system&#x2F;arch&#x2F;expert configs.
评论 #42560270 未加载
评论 #42560217 未加载
mind-blight5 个月前
Super cool idea! What are you doing to ensure consistent results based on the input? E.g.<p>- does the AI perform the same trades given the same input?<p>- does the AI perform the same trades given slightly different inputs? (E.g. same data, but re-ordered)
forgingahead5 个月前
Really cool, you might want to update the main above the fold summary stats to include the unrealised gains, because it looks like nothing is working &#x2F; nothing has happened until you scroll and read around a bit.
datadrivenangel5 个月前
This is fun! What kind of prompts &#x2F; prompting techniques are you using?
评论 #42560034 未加载
lewj5 个月前
I am committed - added to my daily morning reading list! Will be interesting - my gut will state that it will outperform a fair number of ITF&#x27;s, if only due to the inevitable usage by said funds!
jasfi5 个月前
For Gemini you should use either the latest experimental model (gemini-exp-1206) which should become 2.0 Pro, or 2.0 Flash (a released model). The 1.5 Pro model is way behind.
praveen99205 个月前
I think this shows more of bias of market analysis(text) rather than anything. The reasoning will mostly align with analysis.<p>And also pure randomness of picking the one trade from list of trades
jingojango25 个月前
It would be cool if it had a countdown to 6 am PST next day.
评论 #42560247 未加载
bee_rider5 个月前
GPT’s guess makes the most sense. If you are an AI, invest in a competing AI company. If you are obsoleted, maybe you can buy your way out of being shut off.
ttul5 个月前
I’d love to tune in for updates, but the subscribe button says, “ Failed to send verification email.” This is so cool. Would love to follow along.
评论 #42562201 未加载
dghlsakjg5 个月前
Tried to sign up for emails, but got an error message!
评论 #42561960 未加载
Plasmoid5 个月前
I&#x27;m getting &quot;Failed to send verification email&quot; when I try to sign up for your news letter.<p>So props on doing proper double opt-in for newsletters.
评论 #42561954 未加载
TripleChecker5 个月前
If nothing else, I&#x27;m genuinely curious which performs the best over the long-term.<p>Time to add some side wagers and bet on different models.
mattfrommars5 个月前
&gt; Node.js&#x2F;Lambda backend for AI processing<p>Is this AWS? Why did you pick lambda over say Python code, say in Flask to perform actions?
woollysammoth5 个月前
Sounds like a fun experiment! The overflow-x:hidden on body&#x2F;html is causing weird issues when scrolling (on FF.)
pavel_lishin5 个月前
Where do they get the market news from?
评论 #42560010 未加载
mvdtnz5 个月前
Can&#x27;t verify my email address for the sign-up, it sends me to the domain &quot;undefined&quot;.
评论 #42562340 未加载
Animats5 个月前
This just started, apparently. It will be interesting to see where it is in three months.
评论 #42561488 未加载
jingojango25 个月前
What is meant by 5 dollar stakes? The bought shares reach triple digits in price.
评论 #42560057 未加载
malux855 个月前
It would be so funny if Gemini shorted Google and made a huge profit
asdefghyk5 个月前
What, could go wrong?
评论 #42560035 未加载
inSenCite5 个月前
This should be fun to watch
sgammon5 个月前
&gt; Watch AI bots trade<p>&gt; BOUGHT TLRY
unsupp0rted5 个月前
&gt; Best Performer<p>&gt; AIs are tied<p>Sounds about right
评论 #42560063 未加载
评论 #42560085 未加载
tmaly5 个月前
Any chance you can show the source code for this?<p>Thanks and Happy New Year
jesprenj5 个月前
Right now they are just buying, no one is selling ... interesting.
评论 #42560014 未加载
评论 #42560199 未加载
评论 #42560049 未加载
cedws5 个月前
Now this is interesting. An LLM capable of delivering consistent returns even outside of a bull market would be more of an indicator of AGI to me than any of the benchmarks.
bun_terminator5 个月前
Sir, a second scrollbar just hit the towers
评论 #42560212 未加载
dotancohen5 个月前
<p><pre><code> &gt; Best Performer: AIs are tied &gt; Total Profit: $0.00</code></pre>
评论 #42560042 未加载