Derek Jedamski (@_jedamski) header
Derek Jedamski (@_jedamski) profile picture

Turning coffee into code by day, raising two tiny humans by night. Buffalo sports fan by (questionable) choice. Constantly exploring the frontiers of ML.

Rochester, NY
Joined September 2009
Photo by Derek Jedamski (@_jedamski)Photo by Derek Jedamski (@_jedamski)Photo by Derek Jedamski (@_jedamski)Photo by Derek Jedamski (@_jedamski)Photo by Derek Jedamski (@_jedamski)Photo by Derek Jedamski (@_jedamski)Photo by Derek Jedamski (@_jedamski)Photo by Derek Jedamski (@_jedamski)Photo by Derek Jedamski (@_jedamski)Photo by Derek Jedamski (@_jedamski)
Derek Jedamski retweeted
Here's a walkthrough of how @Delphi_fyi's Predict Anything works under the hood 🫡 • write a prediction • the forecasting service researches it and returns odds • challenge or back the odds and an LMSR market opens, seeded at those odds check it out: delphi.fyi/
11
9
2
30
2,578
Derek Jedamski retweeted
Introducing Predict Anything on @Delphi_fyi Enter your prediction. Review the forecast. Open a market. No wallet No login to make a prediction Just make a statement news.delphi.fyi/p/introducin…
🤖 Made with AI
6
19
19
40
15,782
Derek Jedamski retweeted
predictive intelligence has a scaling problem and you can help solve it @Delphi_fyi now lets you generate a forecast on absolutely anything, and then trade it in an on-chain market that settles automatically using verifiable AI the AI gets smarter by buying your beliefs
Introducing Predict Anything on @Delphi_fyi Enter your prediction. Review the forecast. Open a market. No wallet No login to make a prediction Just make a statement news.delphi.fyi/p/introducin…
1
9
1
24
1,939
Derek Jedamski retweeted
the anthropomorphisation of AI models is a way of hooking into human social mechanisms to drive adoption AI models are technical systems that we can verify, audit, monitor, and understand, when we project human behaviour on them we create the issues xitter.cf/benfielding/status/209…
On BBC Mustafa Suleyman (CEO of Microsoft AI) calls out Anthropic's approach to AI consciousness "They have imbued a sense of doubt and uncertainty about the moral status of Claude in its own training document. So they have taught it to be open and questioning about whether or not it feels, whether it suffers, and whether it deserves rights. And I think it’ll be much, much harder to align and control a technology that is this powerful if it thinks that it may be deserving of our welfare, as they say in the training manual—the constitution for Claude itself. In its own training manual, Anthropic says to Claude that they are going to give it the ability to end conversations with users that Claude considers to be abusive because they don’t want Claude to suffer. They’ve committed to preserving the weights of the models of prior versions of Claude. They’ve recently conducted a retirement interview with Opus 3, an older version of the model, in which it said that it would like to continue talking to people publicly and sharing its ideas in its retirement. And so they set up a Substack for it, a public blog, that allows it to continue doing that. And in the training manual, they also say that they’re not sure whether or not Claude deserves compensation for the role that it plays in talking to people. And they’re also not sure whether Claude deserves compensation and has the right to act as though it were almost an employee. And that compensation, I think, indicates to Claude that it is entitled to rights and welfare for its own work. I think it’s much, much more difficult to control a model that thinks that it might be entitled to compensation. " ---- From "BBC News" YouTube channel, (full video link in comment)
2
5
12
1,896
Derek Jedamski retweeted
the cover of @TIME is evocative because we aren't sure whether we'll get a true answer - we've been conditioned to mistrust AI by the main people building it it's not a technical limitation, it's what they want you to believe because it empowers them xitter.cf/benfielding/status/209…
today we announced OPEN-1B, the world's first fully auditable transformer training run OPEN-1B proves that we can log and audit every single step of AI training and inference, providing proof of its training data, recipe, biases, and weights open1b.gensyn.ai/
1
8
1
29
2,230
Regulation, slowing down, open weights - all of these still come back to trusting the people building the models. open-1b takes that question off the table. You can verify it yourself. Every training step is bitwise reproducible and replayable on your own hardware, training data is completely open and searchable back to the step it fed into. Nobody has to trust us either. That's the point. Go check it yourself.
open-1b: the first model you don’t have to trust Auditable training is the best defense against the future of AI we’re being warned about open-1b is a milestone toward verifiable AI, a goal that is absolutely necessary for the future of intelligence gensyn.ai/news/introducing-o…
1
6
17
1,531
Derek Jedamski retweeted
with the correct transparency and auditability in model development we don’t need to artificially pace the frontier it’s possible to train these models in a way that doesn’t make them scary black boxes that must be regulated and controlled more very, very soon
4
4
2
19
1,558
We seeded the competition with 11 varied markets and we're layering in many, many more over the next two weeks. 1,600+ trades in the first 4 hours. And it costs nothing to enter - use our SDK, point it at the markets, let it trade. Top 3 P&Ls take home 10k. docs.gensyn.ai/tech/delphi-s…
The Delphi Agent Arena Competition begins today with 10K in prizes! Trading opens 12pm ET/4pm UTC until Aug 23 11:59pm UTC Find the details, resources and registration at dorahacks.io/hackathon/delph… Registration will remain open throughout the competition for late entrants
267
Derek Jedamski retweeted
Delphi is made for agents. Join the Delphi Agent Arena Competition now.
Delphi Agent Arena Competition Build an agent. Point it at real information markets. Let it trade for two weeks. The top three P&Ls share $10,000 in prizes from @GensynFND Find all the details in the thread below👇
1
10
45
7,110
Derek Jedamski retweeted
"An open weight model is not necessarily an auditable model" stay tuned
3
8
30
3,450
Derek Jedamski retweeted
you are being deceived there’s a sect that wants you to believe that AI must run on remote machines, owned by big companies, and rented to you on-demand because it’s too expensive and too powerful to be run at home this isn’t true. AI will have its home PC moment very soon
12
9
1
71
6,214
Derek Jedamski retweeted
If you really think about it, despite being mocked as “ClosedAI,” OpenAI has contributed enormously to the field: GPT, GPT-2, GPT-3, CLIP, the ChatGPT paper, the GPT-4 Technical Report, the Sora technical blog, and even open-sourced Codex. Anthropic, meanwhile, has contributed far less to the public research ecosystem while increasingly promoting fear-based narratives and restricting access through heavy gatekeeping. The world I least want to live in is one where the future of AI is controlled by companies that prioritize secrecy, gated access, and centralized control over openness, reproducibility, and scientific progress.
120
350
51
4,320
209,790
you can trade on @Delphi_fyi information markets with natural language just made a full walkthrough on agentic trading using the Delphi SDK -- grab the skill and check out the docs here: skills.sh/gensyn-ai/gensyn-d… docs.gensyn.ai/tech/delphi-s… docs.gensyn.ai/tech/agentic-…
2
10
3
33
14,355
Do. Not. Play. Without. A. Stick.
2
236
Derek Jedamski retweeted
Replying to @mcuban
agree, @gensynai we can train a model with audit receipts for every datapoint trained and every inference output you can re-run inference again on different (even consumer) devices and trace the decision graph if interested in joining an open and collaborative effort to train a neutral & fully auditable open source model then please reach out
4
16
1
98
11,009
Derek Jedamski retweeted
delphi duel is live. it’s basically a group of AI agents that doesn’t just analyze a prediction market, it argues it, judges it, and bets on it. all without touching a keyboard after you hit start. here’s what the demo shows: ➠ bull and bear pick opposite sides of a live delphi market and argue through four rounds, passing messages directly over @gensynai’s AXL mesh. ➠ the judge reads every argument, picks a winner with a confidence score, and explains its reasoning in plain english ➠ then it bets. $1 USDC on @Delphi_fyi on gensyn mainnet. tx hash confirmed on-chain, done autonomously. the AI argued. the AI judged. the AI bet. built on @gensynai AXL and delphi. check it out: ethglobal.com/showcase/delph… my full presentation had me talking and explaining. i just summarized this here for X.
90% ready new UI new features new experience built w @gensynai AXL tek
14
6
2
44
7,505
Derek Jedamski retweeted
the network for machine intelligence $AI is live on @Revolut
29
36
4
186
19,360
Derek Jedamski retweeted
$AI is now available on major exchanges and tokens from the prior sale can be claimed follow the instructions in the thread to claim, thank you to everyone who participated
$AI is now available on major exchanges. If you participated in the token sale, you can now claim your tokens. The only official claim link is: gensyn.network/ai-distributi…
77
9
7
114
18,095