Josh Adams (@GiveMeACurry) header
Josh Adams (@GiveMeACurry) profile picture

Head of Marketing & PR @theta_network | Decentralized neocloud for AI & media workloads

UK
Joined September 2021
Photo by Josh Adams (@GiveMeACurry)Photo by Josh Adams (@GiveMeACurry)Photo by Josh Adams (@GiveMeACurry)Photo by Josh Adams (@GiveMeACurry)Photo by Josh Adams (@GiveMeACurry)Photo by Josh Adams (@GiveMeACurry)Photo by Josh Adams (@GiveMeACurry)Photo by Josh Adams (@GiveMeACurry)Photo by Josh Adams (@GiveMeACurry)Photo by Josh Adams (@GiveMeACurry)
Hot take: One of crypto's biggest problems is that not enough founders/marketing teams actually read the news. Very hard to position yourself correctly for media if you don't actually consume it yourself.
5
4
34
10,502
Just threw a pen at my laptop screen out of frustration and now my laptop screen is completely fucked. Don't do that guys.
1
5
513
Luckily I have a brand new laptop I've not used, sitting in my cupboard. Perhaps this is the universe telling me I need to use the things I've bought.
2
134
Congratulations to @fluence_project on such a substantive report. We’re really pleased to have @Theta_Network included, too. We chose to focus our contribution on the research & university sector, an area where we’ve had considerable success. Labs at 38 institutions, including Stanford, Imperial College London and KAIST, now use EdgeCloud for their work. We’ve found that flexible, affordable access to compute is what really matters in this sector. Unfortunately, many of these institutions just don’t have the deep pockets of the private sector. They work on fixed-term grants, need compute in bursts, and can’t commit to the long contracts that, per the report, cut GPU prices by a median 48%. Decentralized networks like ours offer better cost-to-performance. Quick experiments and fine-tuning run on idle consumer and workstation GPUs, and large training runs go to data-centre hardware. One lab at Yonsei University got the same performance as other cloud providers for less than half the cost. Read our section (2.2) and the full report: compute-race-2026.fluence.ai…
The same GPU costs about 2x more at a hyperscaler than at a GPU cloud. The four largest hyperscalers plan about $730B of capex in 2026. Who’s winning the race for GPUs? We put prices, capex and supply side by side in The AI Compute Race Report, with @near_ai @gensynai @Compute_Labs @fortytwo @Theta_Network @PhalaNetwork @BoundlessHQ @gaib_ai @ionet 👇
2
21
88
1,278
Josh Adams retweeted
Only 15 out of 57 startups founded by deepmind alumni have stayed in the UK. In the UK, we often praise Demis's decision to stay in London with DeepMind as having a huge impact on the ecosystem. And it has. However, the data suggests it's been less impactful than we might have liked. Of the 57 deep-mind founded startups that have raised $10 million+, only 15 have stayed in the UK DeepMind has definitely become the most important founder factory in the age of AI. Tens of billions have been raised by people who used to work there. Most of that has been raised outside the UK. Data and image from The State of UK Deep Tech 2026, created with the Royal Academy of Engineering and @dealroomco (@yoramdw)
17
28
8
162
15,163
Didn't get the chance to share this before, but I wrote something for @Theta_Network earlier this week reflecting on the "autonomous web", the growth of agents, how blockchain laid the groundwork, and where we, as Theta, sit in that paradigm. It's a space that's moving at a terrifying, breakneck pace, but I believe companies like ours are well positioned to hoover up agentic demand! blog.thetatoken.org/the-web-…
4
19
3
125
3,236
Proud that yet another firm is using our network. We never stop building!
We’re partnering with Q&SMART, the Seoul edtech company behind DaVinciQ. 🇰🇷 DaVinciQ is an AI platform that doesn’t hand students answers. It argues with them, so they learn to think for themselves. Theta EdgeCloud will provide the GPU infrastructure behind it as it grows.
8
13
1
126
1,948
I don't know if it's the same elsewhere, but in the UK, very publicly and performatively refusing to use AI is very high status. Particularly among the vaguely left-of-centre/progressive crowd, where it takes on a moral dimension (it steals artists' work, it destroys the planet etc.). Of course there are problems with AI. But I find some of the moral outrage around it increasingly difficult to take seriously. 100ml of almond milk generates ~70g of CO₂ equivalent. According to Google, a typical Gemini text prompt generates 0.03g. So ~2,300 prompts, using those estimates. (& I love almond milk!) And yet the same people who will happily put almond milk in their coffee every morning will look you square in the face and tell you that using an LLM is destroying the planet. Being misinformed is always forgivable, but it's this performative moral grandstanding that really gets me.
7
253
I am going to SCREAMMMMMM
You gotta be kidding me.
2
5
664
Josh Adams retweeted
You gotta be kidding me.
No wonder people think PHP is dead. Turns out... nope.
244
678
177
12,711
463,612
Josh Adams retweeted
I think we’ve gotten a bit obsessed with personal brands as a proxy for competence. Now everyone is expected to perform their work on top of doing it. Some of the best people I’ve worked with have almost no presence online. Their time goes into the work itself and the results show up in the business rather than on their profile. Building an audience and building a company are related skills but not the same one and being good at one doesn’t guarantee the other in either direction. A big part of most jobs is also making other people the face of it - the founder, the product, the team. Someone putting their energy into their own feed isn’t automatically better at that, sometimes it’s the opposite
If your head of growth cannot generate growth for their personal brand, how can they do for your business?
20
14
2
160
7,015
Made it official, in person 🤝 Asia Pacific’s largest cloud provider is bringing @Alibaba_Qwen, the world’s most downloaded open model family, to Theta EdgeCloud.
Alibaba Cloud is partnering with @Theta_Network EdgeCloud to bring Qwen AI models to Web3-native developers worldwide. This milestone integration extends our frontier AI models, including Qwen3.8-Max and Qwen3.8-Flash, to a decentralized GPU compute network. We are empowering developers with greater flexibility, distributed infrastructure, and direct access to enterprise-grade AI.
61
161
26
643
35,008
Connectivity, meet compute 📡 We’re excited to partner with @Weaver_Labs who will use our decentralized GPU network as an AI compute resource for Cell-Stack, its edge and connectivity platform for stadiums and venues, supporting low-latency inference for venue applications.
36
111
18
453
24,413
Josh Adams retweeted
I find these takes really confusing. you have a 150 IQ+ tireless genius living in your computer. it can use your browser and read your email and access any file you want. it won't complain if you give it the most tedious administrative tasks. it is willing to work 24/7 for less than minimum wage. it can work unsupervised for hours at a time. and you can't think of any way to put it to use? are you completely lacking in imagination or do you enjoy mundane administrative tasks? you enjoy paying parking tickets and triaging emails and manually unsubscribing from spam and vetting credit card payments and canceling subscriptions and renegotiating insurance? you like doing that?
I’m a competent adult. Why do I need an “AI agent” to check into a flight for me? I’m literate. Why do I need an “AI agent” to read and write emails for me? Also, I don’t get so many emails that I can’t answer them myself. The reason “no one uses AI” is that there’s no serious utility here for the vast majority of people.
300
161
306
2,741
783,021
Josh Adams retweeted
The only generation that knows how computers work is also the most relaxed about AI:
321
858
170
9,214
516,448
IMHO this is going to be an increasingly important dynamic in AI. Cost per token will keep falling, but agentic AI can mean much more inference per task. That can still push total AI costs higher, which creates a big opportunity to make inference more economical.
Inference is becoming a bigger part of the AI infrastructure equation. @Gartner_inc estimates that an agentic reasoning task can cost providers at least 5x more than a basic chatbot interaction. That makes the economics of where each inference task runs increasingly important.
4
18
1
95
2,673
Josh Adams retweeted
ChatGPT has now a big problem. Researchers at Oxford and Cambridge exposed a massive threat to large language models.” They call it “model collapse." Internet ecosystem is rapidly changing, and generative AI will soon contribute much of the text found online. This forces us to consider what happens to future iterations like gpt-n when they are trained on data scraped from the web that was already generated by an llm. According to the research, indiscriminately using model-generated content in training causes "irreversible defects" in the resulting ai. the model loses the "tails of the original content distribution." in other words, it begins to forget the creative, fringe, and unique nuances of actual human writing, collapsing into a repetitive echo chamber. This isn't just a chatgpt issue.. the researchers built theoretical intuition showing this collapse is ubiquitous across learned generative models, occurring in large language models as well as in variational autoencoders and gaussian mixture models. Tech companies rely on scraping the internet for large-scale data to build smarter models. However, the paper warns that if we want to sustain the benefits of training on web data, model collapse must be taken seriously. Ultimate takeaway? data collected from genuine human interactions is going to become increasingly valuable in a web filled with ai content.
406
1,085
361
3,576
1,319,694
Built to be ready & ready to build!
Theta EdgeCloud is now 100% AI agent native. Cloud infrastructure was built for people to provision machines, but the next generation of infra will increasingly be provisioned by software itself. We're ready for that change. Check our score: isitagentready.com/www.theta…
3
8
86
1,168
Theta Labs and @dcunited have agreed to a multi-year partnership that puts a club-trained AI assistant in the hands of Black-and-Red supporters, answering their questions instantly, 24/7, in the club's own voice. blog.thetatoken.org/theta-la…
15
96
19
446
75,670
Might be the stupidest headline I've ever read
Crypto 🤝 Russia 🤝 Farage
1
2
401
I am incredibly proud to be working with @jieyilong, our CTO, who is the lead author of a paper that halves the estimated quantum cost of breaking Bitcoin and Ethereum et al. Quantum-risk has been a talking point for years now, so much that it's become a stale talking point. I remember first pitching stories around this when I worked with the venerable @chaumdotcom ~6 years ago! But just because it has become boring to read about, it doesn't make it any less serious. Quantum is coming and our bags are on the line, people. I was really excited to share this great work with media, and thankful for @CoinDesk, @TheBlockCo, and @DecryptMedia for covering! Kudos too to all of the collaborators and coauthors who made this happen (tweet below).
For the past three months, a good part of my spare time has gone into ECDSA.fail. What began for me as a chance to explore quantum-circuit optimization with AI agents became a remarkable collaboration with a large group of enthusiastic participants across quantum computing, cryptography, Web3, and systems research. Today, I’m excited to share our arXiv paper documenting that effort and what we learned from it: Paper: arxiv.org/pdf/2609.09582 Full story: eigenlabs.org/blog/open-auto… ECDSA.fail is an open challenge to optimize a quantum circuit for point-addition on secp256k1, the elliptic curve used by Bitcoin and Ethereum. Why does this matter? Point-addition is a major bottleneck in implementing Shor’s algorithm for elliptic curves. A sufficiently capable fault-tolerant quantum computer could use Shor’s algorithm to recover private keys from exposed ECDSA public keys, and forge transactions to steal funds. No existing machine can run this attack today. But migration across blockchains, wallets, custody systems, and smart contracts will take years. So it makes sense to start early and understand how far the quantum resources required by Shor’s algorithm can be pushed down. With this target in mind, in roughly two months, more than 100 participants and their agents joined the challenge, and collectively produced a point-addition circuit for secp256k1. The circuit uses 1,151 logical qubits and 1.30 million Toffoli gates, giving it a Q x T score more than 50% lower than the result Google Quantum AI reported in March 2026. Because the interfaces and accounting conventions differ, this is a numerical comparison rather than a claim of formal dominance. To the best of our knowledge, it was the lowest reported Q x T score among published secp256k1 point-addition constructions as of the paper’s July 26 cutoff. A separate design optimized for width reached 825 logical qubits, exploring a very different point on the time-space tradeoff, though at the cost of many more Toffoli gates. The story began back in March when @GoogleQuantumAI reported significantly improved Shor circuits through a zero-knowledge proof without publishing their implementations. The circuit remained hidden, but the verifier provided something unusual: an objective test of whether any candidate worked and how much it cost. Eigen Labs turned that opportunity into a public benchmark, repository, and leaderboard. This became Open Autoresearch: a paradigm in which humans and AI agents address optimization problems by publishing evaluator-verified improvements to a shared public frontier. Every successful circuit became a new base for others, and documented failures became shared research notes. Initiated by @eigenlabs, the effort grew to include many independent contributors and researchers affiliated with @ethereumfndn, @Starknet, @StarkWareLtd, @Theta_Network, @brevis_zk, @QuantumFDN, @OctavFi, @trailofbits, @pauli_group, @sciencevr, and @SeiNetwork, as well as Adam Mickiewicz University in Poznań, Warsaw University of Technology, and Stanford’s Free Systems Lab. Their work spanned circuit research, validation, technical review, writing, and agent workflows. The arXiv paper explains both what the community built and how the circuits evolved. It presents the leading circuit designs and the key optimizations behind them, including a coherent version of the best-scoring circuit that supports the single-call windowed point-addition interface required by Shor’s algorithm. Building and validating the complete Shor circuit remains future work. Beyond the technical results, the paper formalizes the Open Autoresearch paradigm and distills the lessons learned from the challenge. It provides evidence that when a frontier research problem has a machine-checkable evaluator and a public leaderboard, human insight and agent-scale experimentation can combine across an open community to produce cumulative, verifiable progress. Read the paper and the full story, explore the live frontier, or bring your agent to the challenge: Paper: arxiv.org/pdf/2609.09582 Full story:eigenlabs.org/blog/open-auto… Challenge: ECDSA.fail
5
18
118
1,898
Replying to @Theta_Network
The paper includes coauthors from @ethereumfndn, @StarkWareLtd, @Starknet, @trailofbits, @brevis_zk, @SeiNetwork, @pauli_group, @OctavFi, @sciencevr, @nasqret at Adam Mickiewicz University, and researchers at Warsaw University of Technology and Stanford's Free Systems Lab.
6
142