Every year, I put together a year's review of what's going on in the crazily evolving world of brain chips, #Neuralink and #BCI's.
2025 was huge for BCIs, but we saw many other trends.
Here are my top-10 trends of the year and what to watch for in 2026:
#neuroskyence #NeuroAI #NeuroTech
True! First was language, second motor AFAIK?
Give a follow so I can keep you updated about this process.
If you can, please like or repost the first post to support the growth of the brain chips community.
I hope you found this thread helpful.
In 2026, I'll keep updating you on the world of brain chips and expand operations with informative posts about top brain understanding research.
In the long term, I am dedicated to understanding the brain to help improve brain chips.
not merely temporary experiments. They are poised to become reliable, lifelong solutions.
Thanks for this, folks.
x.com/WillettNeur...
#1 Patients and Participants:
2025 was undoubtedly the year of patients.
With Neuralink nearing 20 participants, and leveraging on 20 years of industry experience (14 BrainGate participants, plus at least 20 others), brain-computer interfaces are beginning to prove they're ...
#2 Paradromics:
Paradromics really stole the show this year.
With NEOM funding, great communications, and securing IDE approval for the Connexus BCI, Paradromics proves that high-data-rate implants are moving steadily long-term.
www.linkedin.com/feed/update...
#3 Neuralink's Employees:
We wouldn't be here if it weren't for Neuralink's talent.
Not only have they expanded their clinical trial massively, but also, their 2025 tour shows the industry is now focused on building a long-term force to scale operations.
www.youtube.com/watch?v=QJd...
#4 Science Corp:
Science surprised this year with their light-speed clinical trials, and demonstrated how close they are to commercial availability.
Science Corp's Vision Patient Registry is now open!
science.xyz/news/scienc...
#5 Mainstream Acceptance as "Transformative Medicine":
The "sci-fi" stigma evaporated in 2025.
As reported by TIME, the medical establishment and the public now officially recognize BCIs as a real, life-changing pillar of future medical treatment.
time.com/7330887/bra...
#6 The Renaissance of Dexterous Robotic Hardware:
Brain signals are only half the battle; 2025 was the year the hardware caught up.
Mehrdad Farimani’s mapping of five-finger robotic hands shows how humanoids are pushing the dexterity frontiers.
www.linkedin.com/posts/hmorv...
#7 Global Expansion of BCI Clinical Trials:
We saw more clinical trials move beyond the US this year.
The launch of the UAE-PRIME trial in Abu Dhabi marks the beginning of a global race to bring thought-controlled devices to patient populations.
www.linkedin.com/posts/neura...
#8 The Convergence of AI and Neuroscience:
The lines between "software" and "brain" continued to blur in 2025.
Patrick’s report from NeurIPS shows how the AI community and neuro-labs are beginning to working as one cohesive ecosystem.
www.linkedin.com/posts/pmine...
#9 Germany:
Went under the radar but Germany implanted its first BCI participant for motor recovery with penetrating electrodes this year. This is unique in Europe.
Also, Germany's CorTec had its first implant (in a US center.)
www.linkedin.com/posts/mirmi...
#10 Massive Capital Influx:
2025 saw unprecedented investment confidence, headlined by Neuralink’s $650M Series E.
With ARK Invest and Sequoia joining in, it’s clear the market now views brain chips as the next major tech platform.
x.com/neuralink/s...
Every year, I put together a year's review of what's going on in the crazily evolving world of brain chips, #Neuralink and #BCI's.
2025 was huge for BCIs, but we saw many other trends.
Here are my top-10 trends of the year and what to watch for in 2026:
#neuroskyence #NeuroAI #NeuroTech
😂
The new Titans architecture shows how badly Google wants to go back to recurrence.
Why? Recurrence is damn cheap.
This paper shows how the brain does it.
Striatum, cortex, HPC, BG, and thalamus, in recurrent loops with shortcuts.
A🧵with my toy model and notes:
#neuroskyence #compneuro #NeuroAI
Disclaimers:
1) This is a simplification from a very complex paper. Sorry if I omitted some details, it is all for clarity 😸
2) As any paper, this study has to be further REPLICATED to know if its true.
Link to the paper:
doi.org/10.1016/j.t...
Link to the Titans blog post:
research.google/blog/titans...
Limitations:
The precise anatomical routes for these shortcuts still need experimental verification.
It also remains unclear exactly how the brain governs the reliance on these shortcuts.
Implications:
This reframes addiction and disorders not as broken systems, but as network pathologies.
It also suggests energy-efficient AI could be built using similar "lazy" shortcut architectures.
Finally, they propose behavior isn't binary but exists on a dynamic continuum.
The degree of shortcut usage determines where an action falls between goal and habit (Box 1).
They identified the hippocampus as the generator of potential future sequences.
These "preplay" sequences feed into the loops to initiate goal selection (Figure 1A).
They claim striatal hubs integrate context like AI self-attention mechanisms do.
Converging inputs allow the striatum to create new meaning from scattered signals (Figure 2B-D).
They proposed that habits arise when connections bypass the slow cognitive loops.
Their model showed these shortcuts connect limbic inputs directly to motor outputs (Figure 1B-D).
This is how we learned that from the paper:
They argue that limbic, cognitive, and motor loops are not isolated tracks.
Instead, they share information through mutual crosstalk and overlapping terminal fields (Figure 1A).
That's the gist of it.
But why?
Previous dual-system models failed to explain how goals seamlessly blend with habits.
They investigated this to reconcile behavioral flexibility with the rigid anatomy of known circuits.
Here is another way to put it: the striatum acts like a biological search engine.
It weighs massive inputs to find relevant context, just like "self-attention" mechanisms in modern AI.
What's the brain secret weapon?
The striatum (ST).
ST integration hubs 🧠 do the cross-talk, analog to the self-attention mechanism🤖.
Both systems weigh converging context inputs to extract relevant meaning 💡.
But Brain does it in a single loop instead of 100x layers.
