Peter Christie (@pacc188) header
Peter Christie (@pacc188) profile picture

ASX:ATV Cloud-enabling Telco and DC’s. Building GPU tech for cloud gaming and AI. Perth. Oasis at the corner of the world.

Perth, Western Australia
Joined April 2010
Photo by Peter Christie (@pacc188)Photo by Peter Christie (@pacc188)Photo by Peter Christie (@pacc188)Photo by Peter Christie (@pacc188)Photo by Peter Christie (@pacc188)Photo by Peter Christie (@pacc188)Photo by Peter Christie (@pacc188)Photo by Peter Christie (@pacc188)Photo by Peter Christie (@pacc188)
Peter Christie retweeted
Does agentic AI need a new kind of CPU? Our VP of Product and Core Tech, @aton2006, told @FierceNetwork_ it doesn’t. Agents are asking CPUs to do what computers have always done (query databases, compile code, run tools) just a lot more of it. Read the full piece: bit.ly/4iEyci3
1
5
9
557
Punting the resurgence of the Bachelor's Handbag.
The only green on my screen. Chickens!
2
2
226
NVIDIA's Neocloud customers are yesterday’s Bitcoin miners, CoreWeave, IREN, Firmus, Sharon AI and the outcome looks eerily similar. Both times, GPUs became scarce, memory prices peaked and power became the bottleneck. Back then crypto was going to replace fiat, now AI is replacing labour and both times, the market piled in at the top. Nvidia was up 1,340% at its 2018 crypto-peak this time it’s up 1,310%. Back then Nvidia dropped 54% as inventory caught up and today it’s off 24% but it might be different...really. My optimistic view: Bitcoin peaked near $20K in 2017. The market flushed, then it popped back to $126K in 2025, up 530%, on the back of a software tool with limited utility. AI has real general utility because it does work people get paid for. If a coin with no cash flows could reset and rally 6× past its bubble high, AI is likely to shake out the hype, reset, then rally even harder. Hardware production might struggle to keep up for long time yet.
2
127
The only green on my screen. Chickens!
1
1
2
325
Peter Christie retweeted
Who did this shit lmao
343
2,047
356
13,730
965,935
Close, the lack of moat means the revenue will be shared amongst 20+ vendors that offer the same result +/- 5%
Very few people will pay Anthropic when they can get 95% of the results for 5% of the price from open source providers. Organisations that are need a much better CFO. (Open to multimillion dollar offers... 😉)
79
Forget dot-com analogies, in crypto-years for AI its 2017
2
108
Peter Christie retweeted
dario begged me in dms to not do agent swarms at home but whatever, this is nemotron 120B A12B running 64 concurrents at a bit over 100 tok/sec each for an aggregate of almost 7K tok/sec on a single GB300 (Asus DGX Station) believe it or not, the agents did NOT magically just start hacking the planet tho. weird.
180
169
24
3,139
155,064
if you are in ai (stocks), pivot to shitcos
2
5
2
50
11,269
Peter Christie retweeted
ANTHROPIC FACES CALL TO DELAY MASSIVE IPO OVER AI SAFETY CONCERNS Anthropic has opened discussions with NYC Comptroller Mark Levine after he urged the company to postpone its planned IPO until it addresses AI safety and governance risks. Levine oversees public pension funds holding more than $325 billion in assets and warned that a major AI safety incident after listing could severely damage investors. Anthropic is reportedly targeting an IPO as early as November, following quarterly revenue of $16.8 billion. The key takeaway: AI safety concerns are emerging as a potential obstacle to one of the largest IPOs in history.
8
23
3
289
20,611
Warren the research bot
1
170
Some things make me laugh more than others....
1
89
These are awesome...
192GB RAM in a box. Ryzen AI Max+ Pro 495. This one by GMKtec. let’s plug it in
63
Peter Christie retweeted
"Chinese models are only good at benchmarks." Prime Intellect just rewrote its coding agent in Rust with GLM-5.3 doing the main work. 2 weeks. 2,000+ agents. 200B+ tokens. And this is the model we have TODAY. Result: the agent is ready for input about 13x faster and uses 83% less memory at startup. The code is open, go read what the model wrote. Qwen4 is already in training. I am waiting for the next DeepSeek. I want to see what Moonshot does after Kimi K3 and what Zhipu does after GLM-5.3. If this generation rewrites real software in 2 weeks, the next one is going to be a revolution. I hope we get the weights again. I am so ready for it.
Over two weeks, Prime Agent orchestrated a swarm of over 2,000 agents to rewrite itself in Rust. With 10,000+ sandboxes, 200B+ GLM-5.3 tokens, and 16,000 agent-to-agent messages, this was our largest test yet of Prime Agent’s multi-agent capabilities and our infrastructure powering these agent swarms. Prime Agent now reaches usable input ~13× faster and uses 83% less startup memory than before.
17
19
1
283
20,623
Peter Christie retweeted
OpenRouter’s monthly token traffic has risen from 2 trillion in January 2025 to 379 trillion as of July 2026, with open-weight models accounting for 75% of token generation. $NVDA $MSFT $META $GOOG $AMZN
40
63
19
479
49,217
Peter Christie retweeted
I’d be buying two @NVIDIAAI DXG Stations and allowing the open source community to access 🚀
If you had a $5M networth. Would you spend $250k on a Porsche?
21
6
128
3,610
Peter Christie retweeted
20
88
6
591
14,342
Great to have Edison on board. Ex Megaport Brisbane native in Europe, Edison knows our market sector well. linkedin.com/in/esmithstubbs…
3
150
Great contract >$1M (conservatively). All-up added $2.5M of new TCV in Q1. Mike Glynn's crushing it.
We’re pleased to announce that Microscan has appointed Activeport to build and launch its new Network-as-a-Service (NaaS) platform. Read the full article: activeport.com.au/microscan-… #NaaS #Microscan #Activeport #NetworkInfrastructure #Automation
1
1
3
160
Peter Christie retweeted
We’re pleased to announce that Microscan has appointed Activeport to build and launch its new Network-as-a-Service (NaaS) platform. Read the full article: activeport.com.au/microscan-… #NaaS #Microscan #Activeport #NetworkInfrastructure #Automation
1
1
2
192