verdverm

27 ms·

841.
by verdverm 3mo ago
I always thought this should be easier by telling it to write each letter one a new line, then count, any token separator ought to suffice, so much so you'd think they'd have trained in this strategy given tokens make individual l
842.
by verdverm 3mo ago
basically the same deal as OC Go, lending credence to token commodification
843.
by verdverm 3mo ago
multi turn sessions, they are typically in the high 90% hit rate across all providers without doing much of anything
844.
by verdverm 3mo ago
There are also US based companies like Fireworks serving up the best open weight models with the compliances we need in US enterprise. Depending on the company, they may offer more/different jurisdictions, EU probably needs a Fireworks
845.
by verdverm 3mo ago
I hope all models going forward come with a dflash drafter so we don't have to train one up separately.
846.
by verdverm 3mo ago
The tool I've been using, llm-compressor, can quant models that do not fit in memory (use the sequential pipeline) https://github.com/vllm-project/llm-compressor my setup to help you on your way: https:/&#x2
847.
by verdverm 3mo ago
I know about it, that's the one I'm referring to. That proposal is insufficient. I was deeply involved in the private data discussions, built a PDS fork / prototype on ReBAC/SpiceDB, and presented it at the first Private
848.
by verdverm 3mo ago
Fable has been significantly guardrail'd to the point it refuses to work on login pages because they have auth related activities (as reported by another HN user). There are many such cases of people reporting Fable unwilling to work o
849.
by verdverm 3mo ago
They have introduced https://agenticresourcediscovery.org/ with many other heavy hitters, which looks quite good without deep diving it. I would not bank on this Gemini Enterprise spin because they are try harding lockin (b
850.
by verdverm 3mo ago
I thought they were the ones that are supposed to be making them for me, and to make that digital coinage for me via the ai autonomous business proxy
851.
by verdverm 3mo ago
ATProto's tbd permission system is insufficient for the granularity needed in enterprise. Many of those features needed (like groups) will have to be built in an app view and by proxy be centralized. ACLs are two generations in the pas
852.
by verdverm 3mo ago
yeah, I was surprised when Sundar said they will release it in a month, since training/evals is not so predictable
853.
by verdverm 3mo ago
have we defined what ai limbo is yet? rumor has it that pro is not so pro in internal evaluations and that there are internal stakeholders holding it up
854.
by verdverm 3mo ago
What about the water breaks for ads? That seemed real Americanization to me
855.
by verdverm 3mo ago
we know that 1. Different people using Ai have different outcomes 2. There is much more to agents than the model 3. Companies are not successful because of the code, look at how many shitty products we endure Can you explain the basis for y
856.
by verdverm 3mo ago
"latest is not a version" ~ Kelsey Hightower
857.
by verdverm 3mo ago
Childhood psychologists are warning about giving Ai to children, very similar to social media and generally screen time. How does this factor into your business?
858.
by verdverm 3mo ago
While working on my PhD, there was one guy in the field that would publish new, better results to Springer every time someone improved upon benchmarks. No code, no data, no conference publications, paywall restricted, unverifiable, always t
859.
by verdverm 3mo ago
Another important questions is: What about the papers that graduate to proper publication? Arxiv is full of pre-prints that anyone can upload.
860.
by verdverm 3mo ago
Third options include - fine-tuning - running in your environment
861.
by verdverm 3mo ago
The point is that for many tasks today, and likely all tasks before long, that the open vs closed will not be a differentiator. There are many open models much better than gemini, yet people still use gemini. It's like picking AWS vs G
862.
by verdverm 3mo ago
How does this comparison sound if we use cloud provider? The Ai is writing code, not executing a startup. The code was never the hard part of startups
863.
by verdverm 3mo ago
yet OpenCode is #1 harness per Artificial Analysis https://artificialanalysis.ai/agents/coding-agents?coding-ag...
864.
by verdverm 3mo ago
I pinged the OC devs to get someone else's PR merged to fix a slop artifact, they did indeed merge external PRs. Like many popular Ai projects, they get 1000s of slop PRs. They are also working on a v2, so fixes to old code are much le
865.
by verdverm 3mo ago
I'm using a number of paid per-search like exa / tavily + SearXNG, wrapped in a cli, often with agents doing deep research-ish
866.
by verdverm 3mo ago
automated posting will not do well here, best to get to know the community first
867.
by verdverm 3mo ago
This is the third time in as many days you have made this post which violates the HN Guidelines. Links to the guidelines, FAQ, and Who's Hiring post can be found at the bottom of most pages. https://news.ycombinator.com&#x2F
868.
by verdverm 3mo ago
Just use the ENV vars, https://kkovacs.eu/opencode-go-with-claude-code/ | https://code.claude.com/docs/en/env-vars You can use a project like GoModel to add a great gateway where you can swap
869.
by verdverm 3mo ago
https://foreignpolicy.com/podcasts/ones-and-tooze/ technically a podcast, occasionally about Ai, learning a lot about economics and history
870.
Xi Jinping Delivers Keynote Speech at World AI Conference 2026 in Shanghai [video](youtube.com)
11 points by verdverm 3mo ago|1 comments

More ›