HNHacker News
TopNewBestAskShowJobs

aimanbenbaha

75 karma · joined March 21, 2024

submissionscomments
aimanbenbaha··on What if no one misses TikTok?
Exactly. I know people who have TikTok as the lifeline of their business and they're rightfully pissed off that it's taken from them on a whim.
aimanbenbaha··on California's future depends on how leaders rebuild after the Los Angeles fires
Thinking only flat areas are ripe for high-density urbanism is still a NIMBY mentality. Where I'm from there were much worse constraints yet the rate of growth by which cities are developed vastly outclasses the YIMBYest cities in the US let alone California.

There should be respect to the families that lost their property. But it's not insulting to suggest that NIMBY development exarcebates disasters like this. Heck if land use was as productive as it should in America most of those families wouldn't have a good chunk of their wealth evaporates as a result of a natural disaster, and rebuilding would've been hella more faster and cheaper.

aimanbenbaha··on TikTok preparing for U.S. shut-off on Sunday
You said a lot of words that are a litmus test proving you almost never use TikTok nowadays. First of all this is not 2020 anymore, you're 5 years late if you think TikTok is still filled with "dances". "Paid influencers" mean nothing in TikTok since everyone has an equal voice and equal shot at virality, see you're still seeing it from the lens of Instagram here.

The NPC live streaming is weird yeah but you cherry picked a trend and then make it about all TikTok. Literally hundreds of trends spawn up in TikTok every month and some of them are damn more authentic than whatever happens in IG reels. Some of the successful original trends even pick up in Instagram or YouTube.

aimanbenbaha··on Training AI models might not need enormous data centres
You're right. My caveat not exactly accurate but I wanted to point out where DisTrO might comes in and why it's relevant here.

I mean it reduces the communication overhead by more orders than DiLoCo.

aimanbenbaha··on Training AI models might not need enormous data centres
This bottleneck right here is why Open Source is presented with a golden plate opportunity to lead the training of cutting edge models.

Federated learning breaks the barrier to entry and expands the ecosystem allowing more participants to share compute and/or datasets for small players to train models.

DiLoCo introduced by Douillard minimizes communication overhead by averaging weight updates. What this article misses though is that despite this, each GPU in the distributed cluster still needs to have enough VRAM to load the entire copy of the model to complete the training process. That's where DisTrO comes in which even reduces further the inter-GPU communication using a decoupling technique (DeMo) that only shares the fast moving parts of the optimizer across the GPU cluster.

>And what if the costs could drop further still? The dream for developers pursuing truly decentralised ai is to drop the need for purpose-built training chips entirely. Measured in teraflops, a count of how many operations a chip can do in a second, one of Nvidia’s most capable chips is roughly as powerful as 300 or so top-end iPhones. But there are a lot more iPhones in the world than gpus. What if they (and other consumer computers) could all be put to work, churning through training runs while their owners sleep?"

This aligns with DisTrO techniques because, according to them it could also allow consumer devices like Desktop Gaming PCs to join the compute cluster and share workloads. Besides there's also an open-source implementation called exo that allows models to be split among idle local devices but it's only limited to inference.

Again might still be relevant since in the article it mentions that DiLoCo was able to make the model respond better when faced with instruction prompts or reasoning questions never encountered during pre-training. And Arthur seems to think test-time training will make his approach become the norm.

sources: DisTrO: https://github.com/NousResearch/DisTrO DeMo: https://arxiv.org/pdf/2411.19870 Exo: https://github.com/exo-explore/exo

aimanbenbaha··on Ending our third party fact-checking program and moving to Community Notes model
As someone who speaks French this made me chuckle
aimanbenbaha··on Ending our third party fact-checking program and moving to Community Notes model
Also didn't know Meta was outsourcing fact-checkers which is a very terrible idea that sponsored a shady economy of ghost workers that were paid pennies for reviewing gore content.

It'll really take a special mind to think Community Notes wasn't a positive feature added to the social network sphere. Musk despite his schtick did very bold things that other platforms wouldn't think of doing, such as open-sourcing the recommendation system or recently suggesting the idea of optimising content with unregretted time spent that will reward healthy content and punish toxic content even if the two had the same number of impressions.

The overtone window is shifting towards a more open speech and less of self-gratifying echo chambers that promoted the toxic cancel culture.

aimanbenbaha··on Ask HN: How to learn marketing and sales as a solo entrepreneur?
I would like to add Programmatic SEO and doing Organic TikTok / Reels if you're in the B2C space. These things compound fast and don't take as much time as long term SEO and can be as effective as paid media with luck and persistence.

Of course it'll not be easy so expect to put in effort and reps.

aimanbenbaha··on AI Is Shattering Hope of a Net Zero Future [video]
If anything AI is pushing the market to re-discover nuclear at scale again.

This video is just another degrowther canard.

aimanbenbaha··on Why we use our own hardware
I don't think they're not just aware. But purely from scaling and distribution perspective it'd be wiser to start on cloud while you're still on the product-market fit phase. Also 'bare metal' requires more on the capex end and with how our corporate tax system is set it's just discouraging to go on this lane first and it'd be better off to spend on acquiring clients.

Also I'd guess a lot of technical founders are more familiar with cloud/server-side than with dealing or delegating sysadmin taks that might require adding members to the team.

aimanbenbaha··on GPT-5 is behind schedule
And this chain of prompts cumulated with the improved CoT reasoner would accrue a lot more enhanced results. More in line with what the coming agentic era promises.
aimanbenbaha··on OpenAI O3 breakthrough high score on ARC-AGI-PUB
Akyürek et al uses test-time compute.
aimanbenbaha··on OpenAI O3 breakthrough high score on ARC-AGI-PUB
Because LLMs are on an off-ramp path towards AGI. A generally intelligent system can brute force its way with just memory.

Once a model recognizes a weakness through reasoning with CoT when posed to a certain problem and gets the agency to adapt to solve that problem that's a precursor towards real AGI capability!

← PreviousPage 2 of 2