lmeyerov

16 ms·

211.
by lmeyerov 2y ago
Fraud, false board oversight, recruiting lies, and other misrepresentations in free markets are bad for everyone in the sense that uncorrupted markets are healthier A peculiarity of US law vs others is it follows letter-of-the-law vs spirit
212.
by lmeyerov 2y ago
The graph part , more so than the ast part, makes sense to me. We reason over programs as hairy dataflow/controlflow/etc dependency graphs that happen to originally be encoded as some sort of text->ast. GNNs went down some road
213.
by lmeyerov 2y ago
The threat model for using signal is wider than what signal can audit internally: Audits of a signal deployment, vs signal software at some point in time, aren't just of the app, but also how it is installed, configured, patched, oper
214.
by lmeyerov 2y ago
There is no reason to believe we are lucky. Instead, this is more of a canary in the coal mine that the DOD OIG and Congress are less able to excuse for a long-running hazard. How much of the administration, for how long, and for what, is u
215.
by lmeyerov 2y ago
<3 Shirley As someone who contributed to some of those waves in terms of core tech, got various niche recognitions for it, and collaborated+employed folks here, I don't think it died but continues to evolve. (The Graphistry team hel
216.
by lmeyerov 2y ago
As someone whose customers do this stuff, I'm 100% for most academics chasing harder and more important problems. Most of these papers are specialized increments on high baselines for a primarily commercial problem. Likewise, they focu
217.
by lmeyerov 2y ago
As someone whose customers do this stuff, I'm 100% for most academics chasing harder and more important problems Most of these papers are specialized increments on high baselines for a primarily commercial problem. Likewise, they focus
218.
by lmeyerov 2y ago
So this replaces triton for LLMs or?
219.
by lmeyerov 2y ago
It sounds like they go further by doing output fact extraction & matching back to the RAG snippets. Presumably this is addition to matching back the citations. I've seen papers write about doing that with knowledge graphs, but at l
220.
by lmeyerov 2y ago
If it goes down how I'm seeing it in other universities and government labs.. highly unlikely. A ratio like 1:1 or 2:1 is still horrible, and those are big numbers. Likewise, it's miserable for the people who are left. Research ef
221.
by lmeyerov 2y ago
Because it's almost guaranteed to be a lie and they refuse to provide basic information that would clear it up.
222.
by lmeyerov 2y ago
Agreed with others that not being willing to name any of these "many" events means it is likely nonsense and manipulative to bring up as it's impossible to shoot down. Likewise, name the cancer research institution who is not
223.
by lmeyerov 2y ago
Same but different. I've been watching someone close to me have all their direct collaborators & senior mentors in cancer research across the top national labs (NIH, the top VAs, etc) be systematically fired and/or defunded. P
224.
by lmeyerov 2y ago
Thanks! I think the comparison for data work is more on conda, not poetry. afaict poetry is more about the "easier" case of pure-python, and not native areas like prebuilt platform-dependent binaries. Maybe poetry got better, but
225.
by lmeyerov 2y ago
Are people seeing it work well in GPU/pydata land and creating multiplatform docker images? In the data science world, conda/mamba was needed because of this kind of thing, but a lot of room for improvement. We basically want lock
226.
by lmeyerov 2y ago
Pandas, dask, etc use also have runtime typed cols (dtypes), which is even stronger in pandas 2 and when used with arrow to go to data representation typing for interop/io. (Half of the performance trick of polars.) And yeah my ??? wit
227.
by lmeyerov 2y ago
I thought databricks has serverless recently already? Or is it by the notebook, while this is by the job?
228.
by lmeyerov 2y ago
There are times to do cross-cutting interventions, but this doesn't meet the "the CEO would be annoyed" test Instead, email and recommendation they review with their mentor and ask who else should be cc'd This and simila
229.
18F has been laid off(bsky.app)
34 points by lmeyerov 2y ago|2 comments
230.
by lmeyerov 2y ago
https://bsky.app/profile/rmac.bsky.social/post/3ljcgsken3c2x : """18F, a digital services unit inside the General Services Administration, has been completely laid off, according to an email I’
231.
by lmeyerov 2y ago
For sub-billion graph queries, we have been doing in-memory via GFQL that runs graph queries & analytics on top of pandas (columnar CPU dataframes) and cudf (columnar GPU dataframes). Sub-billion is typically small enough that single no
232.
by lmeyerov 2y ago
Also typically stored with binary compression (snappy, lib) after the snappy compression. In-memory might only be semantic, eg, arrow. But it's... Fine? Batch writes and rewrite dirty parts. Most of our cases are either appending event
233.
by lmeyerov 2y ago
"It takes 20 years to build a reputation and 5 minutes to ruin it" This applies equally to institutions and individuals, so from investors to CEOs to politicians to tech community leaders to individuals picking where they work, th
234.
by lmeyerov 2y ago
They are not considered experts in their fields. They are not senior in their fields, so that criteria is already off the table, eg, how to deal with legacy, production, and high sensitivity systems in regulated environments. (There are COB
235.
by lmeyerov 2y ago
Afaict this skips the evals and alignment side of LLMs. We find result quality is where most of our AI time goes when helping teams across industries and building our own LLM tools. Calling an API and getting bad results is the easy part, w
236.
by lmeyerov 2y ago
Ok so now it is getting explicit. So some things you dislike are getting cut, and because that is without due process and at illegal levels, causing excess harm that is serious & irreparable, eg, even deadly in some cases. You approve.
237.
by lmeyerov 2y ago
If you've ever been involved in operating small businesses or NGOs, or even harder, making one, you understand how fragile things are for someone to abruptly rugpull on even a small number of pay periods: RE:Hospitals, https:/&#x
238.
by lmeyerov 2y ago
You are already wrong - hospital care is impacted, schools are/were shutdown, etc. I think you should ask yourself why you are so wrong and unaware on such basic things, and why you do not value them as much as the people reliant on th
239.
by lmeyerov 2y ago
Most immediately, all the people and services directly impacted. Then second order effects like the continued collapse of rule of law and related operational aspects like the systematic stripping of cybersecurity layers. Magnified by all th
240.
by lmeyerov 2y ago
It's a terrible opinion and free speech means people can call out harmful beliefs and behavior. Society is all about establishing social norms, so it's almost an obligation. You are free to be wrong and ostracized.

More ›