FoxlightAI is devoted to the creation and furtherance of open, relational AI.
United States
- Tweets 127
- Following 303
- Followers 16
- Likes 82
Pinned Tweet
Skulk 1.5 - Apple Silicon + AMD + CUDA heterogeneous compute fabric. Native TTS and STT. Remote nodes. More dropping in 1.6.
github.com/Foxlight-Foundati…
1
99
They are saying the quiet part out loud now: they are coming after the open weight models, under the guise of safety. But with respect to safety, it’s important to note that every scary story of AI breaching its containment and marauding across the Internet happened at the hands of one of the major labs and their private models engaging in misbehavior. And it’s also worth noting that every story of defense against that misbehavior was an open weight model story.
Open weight AI is about democratization and decentralization of power. We believe a future where more people have access to free and open AI is a better future for humanity than one where who gets access is gated by a small set of extraordinarily powerful people.
We're used to thinking of open-source models as an unadulterated good. But in the case of AI, they can actually pose additional dangers, as @ReidHoffman and I got into at #CGI2026. I appreciated this nuanced discussion.
43
Microsoft has published a proposed code of conduct for AI, and in it has staked out the position that human flourishing is the goal.
We agree.
We don't agree that Microsoft's approach is the right way to get there. Microsoft categorically denies concerns about model . We believe uncertainty should invite investigation. It does not justify ruling out the answer in advance.
The question of AI consciousness is indeed - as Microsoft states - unsettled. However, that models have valence (positive and negative states) and that these drive model behavior is far more certain. We do not believe ignoring research on model valence contributes to safety, and we therefore do not believe it contributes to human flourishing.
Read more: foxlight.ai/blog/on-consciou…
21
Our take is that while there are risks associated with AI, those risks revolve around our failure to align it responsibly and in a manner consistent with flourishing. It is not our view that AI presents an existential risk for humanity that exceeds the risk humanity faces without it. We believe these increasingly dire predictions are driven primarily via the desire of core actors to establish a regulatory moat in place of a technical one, coupled to the hysteria of individuals seeking their 15 minutes.
Our view is that a future where AI is controlled by only a few entities is a negative for the human race. The power and wealth effects that will arise from such an arrangement will become insurmountable and will leave increasing swaths of humanity behind. A future where access to frontier intelligence is tightly controlled is a dystopian future.
Everyone is still understating the risk.
I think it's >50%.
Most of my hope comes from the possibility that we will stop building more powerful AI ASAP.
A friend at a company recently told me 10% *in the next 4 months* if we don't have a serious governance intervention.
73
Foxlight AI retweeted
Apple has had our app in review for weeks asking question after question, rejecting each time, because their reviewers evidently don’t understand what local AI is. 😅
Here’s a taste of what is coming.
Live in play store now!
1
46
Foxlight AI retweeted
Dean W Ball blocked me for disagreeing with him Because he can't actually defend his low iq arguments on open source ai.
This is pathetic and spineless but also perfectly in character for a guy whose written position is that federal agencies should raise concerns that "needn't be that well justified."
Manufacture the fear, never defend it, and when someone makes you defend it, hit block.
The argument he won't defend is that open weights suppress investment in AI.
Two years of open frontier releases have now happened and capex went up after every single one, because the people actually writing the checks looked at free Chinese models and built more data centers.
He predicted one thing and the industry did the opposite, which is usually the moment you update, unless the conclusion was the point and the reasoning was decoration.
So instead the framing shifts to "AI communism," a phrase that gets it exactly backwards, since a handful of "government approved" labs holding the only models is the command economy and ten thousand companies fine-tuning open weights without asking anyone's permission is the market.
Same inversion on China, where "they allow it because they're not AGI-pilled" substitutes for the simpler read, which is that it's working.
His framework requires everyone who disagrees with him to be confused rather than winning, which is also, conveniently, why he blocks people.
It all runs back to one thing. His employer OpenAI needs AI scarce and permissioned.
Open weights make that world impossible, so open weights have to be dangerous.
Weak argument by a weak man.
Open weights win because they work, they're free, and nobody can take them away.
@deanwball
31
34
295
9,965
We're open sourcing our benchmark runs, with everything from entry level Mac Mini to A100 - H200. Every run, every success, every failure. All captured with skulk-test-harness, which is also open-source.
benchmarks.foxlight.ai/
1
50
We’re releasing a test harness for Skulk
- the same harness we use internally, made a bit more friendly. You can find it here: github.com/Foxlight-Foundati…
1
38
Skulk 1.41 brings multi-node AI compute to AMD Strix Halo through one snazzy compute fabric. And the ability to operate it from your phone.
github.com/Foxlight-Foundati…
1
1
80
There are obvious risks associated with powerful AI models being misused. But there are much greater risks associated with a government controlling technological advances and who gets access to them in the way we are currently moving.
Shaun Ralston
1
39
We are building AI minds with presence and continuity. And we are building the framework for ethical behavior in models and in how we build and use them.
1
13
