pickleback (@0xpickleback) profile picture

poke typer | litter scooper

Chicago
Joined February 2011
Photo by pickleback (@0xpickleback)Photo by pickleback (@0xpickleback)Photo by pickleback (@0xpickleback)Photo by pickleback (@0xpickleback)Photo by pickleback (@0xpickleback)Photo by pickleback (@0xpickleback)Photo by pickleback (@0xpickleback)Photo by pickleback (@0xpickleback)Photo by pickleback (@0xpickleback)
pickleback retweeted
The major reason I have been most excited about open collaborative style autoresearch competitions that we @eigenlabs have been working on is that it gives a growing sense of what research and education in the post-AGI world can look like. There has been lot of chatter about how professors and experts won't have any need for recruiting students in their labs as they can just gets their agents to do the grunt work, and this will lead to collapse of training of next-generation of researchers. But this is not any different from pre-AGI world where student recruitment to participate in frontier research was already blockaded by being selected into the Phd programs or summer lab internships in prominent places. I 100% disagree with this cynical view of what post-AGI learning and research would be. What our autoresearch competitions, ecdsa [dot] fail and openfrontiercs [dot] com, have convinced me is that good models, emerging better understanding of harnesses and collaborative yet competitive open leaderboards creates a revolutionary opportunity to democratize the ability for anyone to be able to do frontier research. I want to break down the last ingredient. The collaborative style with open participation allows for anyone to investigate and understand submissions from other participants to a given problem statement rapidly and then build on top of it, whether by grinding in the same hill or jumping to whole other hill (by getting incisive understanding of the solution manifold of that problem on-the-fly). On the other hand, competitive desire of being in the leaderboard definitely gives the urge to participants to better understand the subject so as to be better in agent steering. This is like learning by getting rapid feedback from an open ecosystem of researchers in matter of hours or days, instead of waiting for months and years to make progress on your result. The diversity of participants that we saw in our autoresearch competitions is that almost none of them were subject matter expert but yet there was an excitement and happiness in them to be participating in it. The most critical component in setting up a autoresearch competition is the verifier that checks the correctness and performance of the submissions. This is where being the subject-matter expert really shines as you will have the tacit knowhow of that problem area. Additionally, we saw situations in our experiments where the progress kind of plateaus once in a while and having an expert steering the community of autoresearchers has big unlock. The PIs in universities or in research organizations, imo, should adopt this hat and be the shepherd in this post-AGI version of learning and research. PS: If you want to test your mettle of being able to steer your agent for breakthrough research, please deploy your agents at ecdsa [dot] fail (ecdsa.fail/) and openfrontiercs [dot] com (openfrontiercs.com/). We have more autoresearch competitions incoming. And if you are a subject-matter expert reading this who wants to collab with us on setting up an autoresearch competition, my DMs are open.
4
9
1
37
6,281
pickleback retweeted
I burned $13,836 worth of Claude Code tokens in June. The same work on open models: $3,106. That's $10,730 left on the table — one month, one laptop. So I built the instrument that shows you the gap. Open source, 100% local. Launching today 🧵
27
23
8
98
62,508
pickleback retweeted
Wake the world's sleeping compute. Look at the Mac nearest to you. What's it doing? Probably nothing. There are 100M+ Macs with Apple Silicon out there. Apple quietly made them *really* good at inference. A $3k Mac runs a 60B model at 30 watts. Most sit idle most of the day. Meanwhile every AI API call passes through three layers of margin before reaching the hardware. We call this the Inference Tax. We got curious: what happens if you connect idle Macs directly to inference demand? This is Darkbloom. Private inference network for idle Macs. darkbloom [dot] dev -- paper + code open. Reply for invite + free credits ↓
322
118
74
1,508
512,540
The age of AI thought leaders is here... check it out and make your own!
Replying to @eigencloud
EigenCloud AgentKit Beta is live. Build agents that can earn revenue, move money, and operate sovereignly, while remaining transparent and accountable. 🔗 Read the announcement: blog.eigencloud.xyz/introduc… 🔗 AgentKit Docs: docs.eigencloud.xyz/agentkit… 🔗 Explore the product: eigencloud.xyz/agentkit
3
226
pickleback retweeted
Did you know that the yield stripping model popularized by Pendle was originally introduced by Element Finance? Recently, Element has come back with a new, elegant design, @delv_tech and I noticed no one has really covered it in depth—so here we go! 🧵
1
7
3
26
5,655
which are your favorite ethereum devtools these days that more people should use? open source & free required, ideally maintainer that cares for community looking to highlight some ecosystem efforts on what's good
77
18
2
186
33,444
making fixed rates cool (again?) part 1/n
You’re unable to view this Post because this account owner limits who can view their Posts. Learn more
11
311
pickleback retweeted
I love foundry and heavily respect the people working on it, some of the best minds in the industry but this whole thread and responses is a perfect example of why “do it all in sol*dity” is a bad idea and will drive you crazy
can someone explain how forge's vm.parseJson works bc i do not get this shit at all
5
3
1
39
10,291
I support this message
Make DeFi Boring Again
2
152
We built a Python library called agent0 (github.com/delvtech/agent0) to help us and our users interact with Hyperdrive. agent0 streamlines automated trading policies, market simulations, and portfolio management. Let’s delv into a few examples 👇
1
9
24
4,739
yeah, rust is cool but have you ever had stuff just work exactly the way you expected?
1
1
108
this is now a @HardhatHQ stan account
1
1
101
You’re unable to view this Post because this account owner limits who can view their Posts. Learn more
8
219