MViking τ (@MViking86) profile picture

Decentralized AI is the only way. Small time crypto gem hunter. Life long New England Patriots fan. Way too into Fantasy football.

Joined January 2024
Photo by MViking τ (@MViking86)Photo by MViking τ (@MViking86)Photo by MViking τ (@MViking86)Photo by MViking τ (@MViking86)Photo by MViking τ (@MViking86)Photo by MViking τ (@MViking86)Photo by MViking τ (@MViking86)Photo by MViking τ (@MViking86)Photo by MViking τ (@MViking86)Photo by MViking τ (@MViking86)
Travesty
The NFL thinks Diggs is worth $7.15M on a 1 yr deal and Boutte is worth a backup safety and 2028 R7 pick. Maye went 14-3 put up 35 TDs, 4,844 total yds, and led the league in passer rating & completion% with those two as his TOP receivers and didn’t win MVP. 🧐
59
Picked up an orphaned team with CMC and got this trade done the first hour. 2027 1st is projected early (3rd) and 2027 2nd is projected late (11). 12 team PPR dynasty- How did I do? #fantasyfootball #dynasty
1
81
Made too many trades to keep track of and have an entirely new lineup lol. What do you guys think, can I take home the ship? #JRPDYNASTYEAST1 @FourPatsDynasty
1
108
One of the big trades I made, I gave up Bijan/Goff for picks side
1
56
Big trade I made shortly after the above, sent most of the picks I received for Hampton/KC.
57
Draft is far from over but here’s what my first several picks look like… Tried to take advantage of the TEP scoring, what do you guys think? Am I cooking or cooked? #JRPDYNASTYEAST1 @FourPatsDynasty
275
With my first pick of the draft, I went with Drake Maye. Very easy decision for me as a lifelong Patriots fan! #JRPDYNASTYEAST1 @FourPatsDynasty
1
272
Got in on this new #Dynasty league with a bunch of #fantasyfootball junkies. SF TEP. Draft is today and I am ready to cook with the 2nd pick! Follow along! #JRPDYNASTYEAST1
This Post is from an account that no longer exists. Learn more
7
440
Recently made a highly concentrated bet on $TAO subnet 15 @oroagents. This one feels insanely bullish and destined for top 20 atleast.
127
Really starting to feel like $TAO and $ZEC are going to be the alt leaders and head to the top 5-10. Excited to see how this plays out.
1
394
Makes you wonder indeed $TAO
$tao ripping like this in the midst of the world ending makes you wonder what will happen when things settle down
1
209
I fully expect @basilic_ai and @grail_ai to keep ripping with all the buzz around @tplr_ai … Only a matter of time until everyone realizes they’re all part of the same @covenant_ai stack. $TAO
161
First it was @tplr_ai. Next it was @TargonCompute. @QuasarModels is up next… $TAO
This is Quasar Attention, the mechanism behind the upcoming Quasar models, designed to support context lengths of up to 5 million tokens. Attention has long been a bottleneck for processing extended context. Standard attention mechanisms struggle to scale beyond ~200k tokens in training, creating a ceiling on how much information models can reliably use. One approach to solving this has been linear attention methods, such as gated delta attention (used in Qwen 3.5) or Kimi delta attention. These improve efficiency and allow longer sequences, but introduce trade-offs: instability at extreme lengths, quality degradation, and in practice, they are not strictly linear. Quasar Attention takes a different approach. It uses a continuous-time formulation, implemented as a fully matrix-based system rather than relying on vector-state approximations. In practice, this improves stability, reduces cost, and maintains performance as sequence length increases. In internal stress tests at 50 million tokens, KDA-based approaches begin to lose stability, while Quasar Attention remains stable. This allows performance to hold as sequence length increases, rather than degrading beyond a fixed threshold. On BABILong, a Quasar-based model pretrained on 20B tokens and fine-tuned on 16k sequences was evaluated on contexts ranging from 1 million to 10 million tokens, maintaining consistent performance across that range. By contrast, models using gated delta attention show significant degradation at longer lengths, in some cases dropping to ~10% performance at 10 million tokens. (Note: results are indicative; setups are not directly comparable) On RULER benchmarks, a Quasar-10B model (built on Qwen 3.5 with frozen base weights and Quasar Attention added), pretrained on 200B tokens, achieved 87% at 1 million tokens, outperforming significantly larger baselines, including Qwen3 80B, under the same evaluation conditions. Taken together, this points to a shift in where long-context performance is won or lost: not in model size alone, but in the attention mechanism itself. Quasar Attention represents a step change in long-context modelling, setting a new standard for stability and performance at scale. We thank @TargonCompute for the compute and for being our compute provider and long-term partner in training the upcoming Quasar models Here is the link to our paper 👇
186
Let’s be honest for a second… 99% of crypto is useless vaporware with no value, that will ultimately go to 0. Most of the “legitimate” coins in the top 50-100 are all competing against each other for a piece of the pie that $ETH currently dominates. But you know what is the perfect use case for crypto/blockchain technology and has mostly untapped, infinite potential? That’s right… Decentralized AI. I shouldn’t have to explain why we can’t afford to let OpenAI, Google, Anthropic, etc dominate the AI space. That’s why I choose $TAO and $Qubic. $TAO is the no-brainer, clear cut leader of decentralized AI, destined for the top 5. $Qubic is the under the radar, sub 200 million MC project that is redefining what’s possible in crypto, 14 years in the making by @c___f___b who has been around since the earliest days of $BTC, and is single handedly responsible for several of the biggest breakthroughs in crypto technology. DYOR.
3
4
22
737
Took a calculated risk on the dip and loaded a significant amount of $TAO into $sn15 @oroagents. Currently building in stealth but my research leads me to believe when they launch, this one is going to be massive. ⏳
1
169